Signatures of natural and unnatural selection: evidence


Signatures of natural and unnatural selection: evidence from an 
immune system gene in African buffalo

Lane-deGraaf, K. E., Amish, S. J., Gardipee, F., Jolles, A., Luikart, G., & Ezenwa, 
V. O. (2015). Signatures of natural and unnatural selection: evidence from an 
immune system gene in African buffalo. Conservation Genetics, 16(2), 289-300. 
doi:10.1007/s10592-014-0658-0

10.1007/s10592-014-0658-0

Springer

Version of Record

http://cdss.library.oregonstate.edu/sa-termsofuse

http://survey.az1.qualtrics.com/SE/?SID=SV_8Io4d9aAYR1VgGx
http://cdss.library.oregonstate.edu/sa-termsofuse


Tools and Technology

A Method for Improving the Reliability of
Sound Broadcast Systems Used in Ecological
Research and Management

JONATHON J. VALENTE,1 Department of Forest Ecosystems and Society, Oregon State University, Corvallis, OR 97331, USA

CHRISTA L. LEGRANDE, Department of Environmental and Forest Biology, State University of New York College of Environmental Science and
Forestry, Syracuse, NY 13210, USA

VINCENT M. JOHNSON, Fisheries, Wildlife, and Environmental Studies Department, State University of New York Cobleskill, Cobleskill,
NY 12043, USA

RICHARD A. FISCHER, United States Army Engineer Research and Development Center, Environmental Laboratory, Vicksburg, MS 39180, USA

ABSTRACT Automated sound broadcast systems have been used to address a variety of ecological questions,
and show great potential as a management tool. Such systems need to be reliable because treatments are often
applied in the absence of a human observer and system failure can cause methodological ambiguity. During
the breeding seasons of 2012 and 2013, we used a sound broadcast system previously described by Farrell and
Campomizzi (2011) in an experiment evaluating the use of post-breeding song in forest-bird habitat
selection in southern Indiana, USA. This system incorporates a portable compact disc (CD) player where the
play button is permanently depressed using manual compression so that when a timer connects an electrical
current to the unit, the CD player automatically starts. Despite exhaustive efforts to find a reliable way to
manually compress the play button on numerous CD player models, play button failure was the most
significant source of broadcast system failure (88%) in 2012. We attempted to resolve this problem in 2013 by
removing the need for manual compression and soldering the play button contact poles on each CD players’
integrated circuit boards. Though we did experience broadcast system failures during <5% of treatment
periods in 2013, none of those were attributable to play button failure. By removing all possibility of failure
from manual play button compression we improved our system reliability. Thus, soldering the CD player play
button on such broadcast systems represents a methodological improvement that can be used by researchers
and managers interested in sound broadcast. � 2014 The Wildlife Society.

KEY WORDS bird song, CD player, momentary action pushbutton switch, play button depression mechanism, sound
broadcast system.

Sound broadcast systems serve a variety of purposes in
ecological research. They have been used to answer questions
about perceived predation risk on reproductive performance
(Eggers et al. 2006, Zanette et al. 2011), movement ecology
in fragmented landscapes (Sieving et al. 1996, 2000, 2004;
Desrochers and Hannon 1997; Bélisle and Desrochers 2002),
habitat selection (Doligez et al. 2002, Ward and Schlossberg
2004, Nocera et al. 2006, Fletcher 2007, Betts et al. 2008),
heterospecific interactions (Diego-Rasilla and Luengo 2004,
Fletcher 2007, Pupin et al. 2007), impacts of noise pollution
(Bee and Swanson 2007), and to explore mechanisms driving
species distribution patterns (Fletcher 2009). Additionally,
audio broadcasts can be incorporated into survey protocols to
improve detectability for some organisms (Gibbs and Melvin
1993, Conway and Simon 2003, Kubel and Yahner 2007,
Ichikawa et al. 2009), and manipulation of species responses

to conspecific attraction shows great promise as a manage-
ment tool (Ward and Schlossberg 2004, Ahlering and
Faaborg 2006).
In many instances, application of auditory experimental

or management treatments is done most efficiently using
automated systems that can be programmed to operate
without a human present (Ward and Schlossberg 2004,
Fletcher 2007, 2009; Betts et al. 2008; Zanette et al. 2011).
Drawing correct and useful conclusions from such studies is
contingent on broadcast treatments being applied correctly,
making the reliability of automated systems a critical
consideration. System failures lead to heterogeneity and
ambiguity in treatment applications, and unreliable systems
need to be visited with greater frequency, resulting in a loss of
money and man hours.
Farrell and Campomizzi (2011) presented a description of a

sound broadcast system that incorporated a portable compact
disc (CD) player on which the play button was permanently
depressed so that when a timer switch connected an electrical
current to the unit, the CD player automatically started.
These authors describe using “a combination of adhesive

Received: 30 September 2013; Accepted: 26 April 2014
Published: 18 August 2014

1
E-mail: jonathon.j.valente@gmail.com

Wildlife Society Bulletin 38(4):827–830; 2014; DOI: 10.1002/wsb.468

Valente et al. � Improving Sound Broadcast Systems 827



tape, wooden dowels, and rubber stoppers to set the play
button in the on position” (Farrell and Campomizzi
2011:463). During field tests with comparable broadcast
systems, we attempted to use similar tools for our play button
depression mechanisms (PBDM), yet we were unable to
develop a reliable method of securing the play buttons in the
on position using compression. As such, PBDM failure was
by far our greatest source of system failure, and we sought to
develop a more reliable method of initiating playback for
our experiments by bypassing the need for a PBDM on the
CD player component. As part of a broader experiment
evaluating the use of post-breeding song in breeding habitat
selection by wood thrush (Hylocichla mustelina) and Kentucky
warblers (Oporornis formosus) in forest tracts of southern
Indiana, USA, we evaluated whether soldering the electrical
play button connection would be a more reliable method of
initiating playback in broadcast systems regulated by a timer
than consistently applying pressure to the play button.

STUDY AREA

Our study area encompassed approximately 750,000 ha of
land in the central hardwoods region of southern Indiana,
specifically within Big Oaks National Wildlife Refuge, Naval
Surface Warfare Center Crane, Martin State Forest, and the
Lost River Tract of Hoosier National Forest. This region
was dominated by corn and soybean agriculture and remnant
tracts of temperate broadleaf and mixed forests. The mean
annual rainfall in southern Indiana was approximately
119 cm (Indiana State Climate Office 2002), and mean
annual temperatures ranged from 68 C in winter to 188 C in
summer (National Climatic Data Center 2011).

METHODS

During the 2012 and 2013 breeding seasons, we deployed
post-breeding song broadcast systems at point-count stations
(34 in 2012 and 24 in 2013) previously unoccupied by target
species (wood thrush and/or Kentucky warblers). Treatments
were applied between 28 June and 23 August each year in
order to experimentally test whether post-breeding song
would attract future breeders. In ideal circumstances,
treatment sites received between 48 and 56 hours of song
broadcast (8 hr/day) in intervals of 6–7 days based on the life
of the batteries (12 V 12 A Energy Power Absorbent Glass
Mat battery; Energy Battery Group, Atlanta, GA). Our
study design incorporated paired control sites, though
information about those is tangential to our focus and will
not be discussed further.
All broadcast systems contained 1 of 2 portable CD player

models (INSIGNIA NS-P4112 or INSIGNIA NS-P113;
INSIGNIA Products, Richfield, MN). The play buttons
on both CD player models utilize a momentary action
pushbutton switch that, when depressed, connects power
required to play the CD to 4 contact poles on the integrated
circuit board via a convex disk of flexible metal (Fig. 1). In
2012, we rotated 18 broadcast systems among treatment sites
on a bi-weekly schedule, and each time a unit was placed at a
treatment site, a trained technician used a combination of
adhesive tape, rubber bands, metal pins, square keys, sticky

tack, glue, and pieces of eraser to manually depress and secure
the CD player play button. Methods for permanently
depressing the play buttons varied temporally as we
attempted to identify a reliable solution, and varied among
CD players because of subtle differences in model design.
In summer of 2013, we increased the number of broadcast

systems to 24 and did not rotate them among sites. To
eliminate the need to apply manual pressure to CD player
play buttons, we applied a thin layer of electrically conductive
rosin soldering flux (RadioShack Corporation, Fort Worth,
TX) to the contact poles and connected them with solder
using a Weler soldering iron (Cooper Hand Tools, Apex,
NC) prior to system deployment (for detailed instructions,
see Supporting Information, available online at www.
onlinelibrary.wiley.com). This connection allows a constant
electrical current to flow among contact poles to create a
closed contact system, which is electrically identical to having
the play button depressed at all times (Eaton Corporation
2007). Both PBDMs and soldering ensure that the 4 poles
are connected when the CD player receives power, but by
soldering the connection we removed virtually all possibility
of manual failure (e.g., play button deformation, PBDM
dislodging, rubber bands stretching, adhesives failing, or
human error in application). We then re-assembled the CD
players to their original condition and incorporated them as
components in the broadcast systems. All other components
used in the treatments were identical for each year.
We visited treatment sites once every 6 or 7 days to change

the batteries and rotate the units (2012 only), and considered
each period between battery changes (treatment period) as
the experimental unit for our analyses. This is a logical
temporal unit because a broadcast system operator would
ideally be able to deploy a system and be confident that the
components would continue working until the battery was
scheduled to expire. When convenient, we made additional
visits to treatment sites to verify that the units were working
as intended.

Figure 1. A close-up image of the exposed upper integrated circuit board of
a CD player (INSIGNIA NS-P4112; INSIGNIA Products, Richfield,
MN). When the 4 contact poles (indicated by red arrow) associated with the
play button are connected by an electrically conductive medium, the unit
begins playing.

828 Wildlife Society Bulletin � 38(4)

http://www.onlinelibrary.wiley.com/
http://www.onlinelibrary.wiley.com/


If a unit was visited before the battery was scheduled to
expire, and the unit was not playing appropriately, a
technician thoroughly assessed all components, documented
the reason for failure, and fixed the source of failure. When
units were visited after the battery had likely expired, the
technician connected a new, charged battery, and then
assessed the status of the unit. Any broadcast system
malfunctions that could not be explained by a dead battery
were also documented. If the playback unit or one of its
components failed at least once during a treatment period,
whereby bird song was not being broadcasted as intended,
the treatment was considered a failure; whereas, if the unit
continued broadcasting until the battery expired, it was
considered a success. Prior to analyses, we eliminated all
failures resulting from human error (i.e., avoidable situations
in which the broadcast system was not properly assembled by
field personnel) and limit our discussion to treatment periods
that were disrupted by component failure.

RESULTS

Our sample size was 100 treatment periods in 2012, and we
experienced 8 treatment period failures, 7 (88%) of which
were specifically attributable to PBDM failures. All PBDM
failures occurred when rubber bands stretched or snapped,
depression components moved slightly so that pressure was
no longer being applied in the correct area, or the CD player
became warped. The only other failure occurred when the
micro Universal Serial Bus (USB) speaker charger stopped
functioning in one unit. In 2013, our sample size was 170
treatment periods, and we experienced 8 total failures, none
of which were attributable to the CD player play button. All
failures were due to malfunction of other broadcast system
components, including 1 speaker, 3 speaker chargers, 1 12-V
direct current (DC) power outlet y-adapter, and 3 CD player
power jacks. Therefore, the soldering method had a 100%
success rate while units were deployed in the field.

DISCUSSION

In 2012, when all of our playback systems were activated
using manual depression (PBDMs), play button failure was
by far the most significant source of treatment interruption.
In 2013, we were able to completely eliminate this problem
by soldering the contact poles that are connected when the
play button is depressed, resulting in a much more reliable
broadcast system.
We encountered 2 primary issues that made manual play

button depression challenging. First, each CD player model
(we experimented with several additional models before
selecting those we used in the field) has a unique design that
requires a unique solution for depressing the play button,
meaning there is no ubiquitous methodology. Second,
because the play button is designed to spring back after
depression, applying too little pressure means the unit will
not start playing as intended, and applying too much pressure
can result in deformation of multiple pieces (e.g., the plastic
play button, the convex metal disk, and the CD player shell
itself) or physically prevent the CD from turning. In addition
to the materials described above, we tried numerous other

methods for securing CD player play buttons in the
laboratory (including manual clamps, glue, rubber stoppers,
and weighted pressure), but were never able to generate an
infallible solution until we eliminated the need for manual
pressure altogether.
There are 3 explanations for why we saw a greater number

of electronic component failures in the second season. First,
it is possible that by tampering with the internal circuitry on
the CD players, we interfered with other internal mecha-
nisms, causing the DC power jacks to eventually fail on 3 of
them. However, we believe that this scenario is very unlikely,
given that every CD player that we disassembled, soldered,
and put back together (n¼27) worked as intended initially,
and none of them failed until after they had been used in the
field for several weeks. Second, our study region received
more than twice as much precipitation in the summer of
2013 (approx. 32.8 cm) as in the summer of 2012 (approx.
16.4 cm; Indiana State Climate Office 2013). It seems likely
that environmental moisture was responsible for many of the
failures we endured with CD players and other electronic
components in the second year, particularly given that others
have identified moisture as a primary cause of failure in
similar systems (Farrell and Campomizzi 2011). Third,
approximately 75% of the playback unit components
(including 18 of 27 CD players) we used in 2013 had
been previously used in the field in 2012. As such, some may
have simply failed from long, sustained periods of use in
adverse environmental conditions (e.g., high heat and
humidity).
Though we only used 2 different CD player models in our

field tests, our experiences suggest that our method can be
utilized in most portable CD player models, eliminating
the need to design a unique PBDM in each case. For
instance, we employed our soldering method on a third
model (Memorex MD8151SL; Imation Corporation,
Oakdale, MN) that utilizes a 4-pole momentary action
pushbutton switch and on a fourth model (SONY D-EJ011;
SONY Electronics, Inc., San Diego, CA) that utilizes a
right-angle tactile switch. In each case, our method was
successful at initiating playback, yet both of these latter
models incorporated electronic volume buttons that are
connected to the same integrated circuit board as the play
button, rendering volume control unusable when the play
button is activated (either by manual compression or
electrical connection). We, therefore, chose to use the
INSIGNIA models, which had a volume control dial not
reliant on the integrated circuit board associated with the
play button, and other experimenters should take this into
consideration when designing their own systems.
Manipulating animal populations using broadcast systems

for both experimental and management purposes holds great
promise (Nocera et al. 2006, Betts et al. 2008, Ahlering et al.
2010, Farrell et al. 2012), yet can also be expensive, time-
consuming, and logistically challenging to implement.
Drawing correct and useful conclusions from manipulative
experiments or management procedures that involve sound
broadcast is contingent on broadcast treatments being
applied correctly, often in the absence of a human observer.

Valente et al. � Improving Sound Broadcast Systems 829



In such instances when a broadcast system fails, there is no
way to determine when the failure occurred, making it
impossible to quantify the number of broadcast hours for
that treatment period and thereby adding another source of
variability to the experiment. In the systems we used, for
instance, failure of any component could result in between
0 hours and 56 hours of lost playback time, depending on
when during the treatment period the component failed. As
such, any improvement in system reliability will, in turn,
improve our ability to generate and accurately interpret
results, and reduce the number of man hours required to
maintain such experiments. We suggest that our method of
soldering the CD player play button is a reliable complement
to broadcast systems used in ecological research and for
management practices that involve using sound broadcast.

ACKNOWLEDGMENTS

We would like to thank M. Betts, C. Bochmann, K.
McCune, B. Slaby, R. Snowden, J. Suich, A. Tucker,
J. Valente, C. Winter, and the helpful staff from the Best
Buy store in Bloomington, Indiana, for their assistance in
designing reliable broadcast systems. We also thank S.
Andrews, J. Robb, and B. Walker for their logistical
assistance, and S. DeStefano, K. McCune, and two
anonymous reviewers for their help in improving this
manuscript. This research was conducted under contract to
the Department of Defense Strategic Environmental
Research and Development Program. The publication of
this paper does not indicate endorsement by the Department
of Defense (DoD), nor should the contents be construed as
reflecting the official policy or position of the DoD.
Reference herein to any specific commercial product, process,
or service by trade name, trademark, manufacturer, or
otherwise, does not necessarily constitute or imply its
endorsement, recommendation, or favoring by the DoD.

LITERATURE CITED
Ahlering, M. A., D. Arlt, M. G. Betts, R. J. Fletcher, J. J. Nocera, and M. P.
Ward. 2010. Research needs and recommendations for the use of
conspecific-attraction methods in the conservation of migratory songbirds.
The Condor 112:252–264.

Ahlering, M. A., and J. Faaborg. 2006. Avian habitat management meets
conspecific attraction: if you build it, will they come? The Auk 123:
301–312.

Bee, M. A., and E. M. Swanson. 2007. Auditory masking of anuran
advertisement calls by road traffic noise. Animal Behaviour 74:1765–
1776.

Bélisle, M., and A. Desrochers. 2002. Gap-crossing decisions by forest birds:
an empirical basis for parameterizing spatially-explicit, individual-based
models. Landscape Ecology 17:219–231.

Betts, M. G., A. S. Hadley, N. Rodenhouse, and J. J. Nocera. 2008. Social
information trumps vegetation structure in breeding-site selection by
a migrant songbird. Proceedings of the Royal Society B 275:2257–
2263.

Conway, C. J., and J. C. Simon. 2003. Comparison of detection probability
associated with burrowing owl survey methods. Journal of Wildlife
Management 67:501–511.

Desrochers, A., and S. J. Hannon. 1997. Gap crossing decisions by forest
songbirds during the post-fledging period. Conservation Biology
11:1204–1210.

Diego-Rasilla, F. J., and R. M. Luengo. 2004. Heterospecific call
recognition and phonotaxis in the orientation behavior of the marbled
newt, Triturus marmoratus. Behavioral Ecology and Sociobiology 55:
556–560.

Doligez, B., E. Danchin, and J. Clobert. 2002. Public information and
breeding habitat selection in a wild bird population. Science 297:1168–
1170.

Eaton Corporation. 2007. Pushbuttons and stacklights 101 basic training:
training manual. Eaton, Moon Township, Pennsylvania, USA.

Eggers, S., M. Griesser, M. Nystrand, and J. Ekman. 2006. Predation risk
induces changes in nest-site selection and clutch size in the Siberian jay.
Proceedings of the Royal Society B 273:701–706.

Farrell, S. L., and A. J. Campomizzi. 2011. A component system for
broadcasting sound for research and management. Journal of Wildlife
Management 75:463–466.

Farrell, S. L., M. L. Morrison, A. J. Campomizzi, and R. N. Wilkins. 2012.
Conspecific cues and breeding habitat selection in an endangered wood
warbler. Journal of Animal Ecology 81:1056–1064.

Fletcher, R. J. 2007. Species interactions and population density mediate
the use of social cues for habitat selection. Journal of Animal Ecology
76:598–606.

Fletcher, R. J. 2009. Does attraction to conspecifics explain the patch-size
effect? An experimental test. Oikos 118:1139–1147.

Gibbs, J. P., and S. M. Melvin. 1993. Call-response surveys for
monitoring breeding waterbirds. Journal of Wildlife Management 57:27–34.

Ichikawa, K., T. Akamatsu, T. Shinke, K. Sasamori, Y. Miyauchi, Y.
Abe, K. Adulyanukosol, and N. Arai. 2009. Detection probability of
vocalizing dugongs during playback of conspecific calls. The Journal of the
Acoustical Society of America 126:1954–1959.

Indiana State Climate Office. 2002. About Indiana climate. <https://
climate.agry.purdue.edu/climate/narrative.asp>. Accessed 24 Sep 2013.

Indiana State Climate Office. 2013. Weather summary. <https://climate.
agry.purdue.edu/climate/summary.asp>. Accessed 24 Sep 2013.

Kubel, J. E., and R. H. Yahner. 2007. Detection probability of golden-
winged warblers during point counts with and without playback
recordings. Journal of Field Ornithology 78:195–205.

National Climatic Data Center. 2011. NOAA’s 1981–2010 climate
normals. <http://www.ncdc.noaa.gov/oa/climate/normals/usnormals.
html>. Accessed 24 Sep 2013.

Nocera, J. J., G. J. Forbes, and L. A. Giraldeau. 2006. Inadvertent social
information in breeding site selection of natal dispersing birds.
Proceedings of the Royal Society B 273:349–355.

Pupin, F., R. Sacchi, A. Gentilli, P. Galeotti, and M. Fasola. 2007.
Discrimination of toad calls by smooth newts: support for the
heterospecific attraction hypothesis. Animal Behaviour 74:1683–1690.

Sieving, K. E., T. A. Contreras, and K. L. Maute. 2004. Heterospecific
facilitation of forest-boundary crossing by mobbing understory birds in
north-central Florida. The Auk 121:738–751.

Sieving, K. E., M. F. Willson, and T. L. De Santo. 1996. Habitat barriers to
movement of understory birds in fragmented south-temperate rainforest.
The Auk 113:944–949.

Sieving, K. E., M. F. Willson, and T. L. De Santo. 2000. Defining corridor
functions for endemic birds in fragmented south-temperate rainforest.
Conservation Biology 14:1120–1132.

Ward, M. P., and S. Schlossberg. 2004. Conspecific attraction and the
conservation of territorial songbirds. Conservation Biology 18:519–525.

Zanette, L. Y., A. F. White, M. C. Allen, and M. Clinchy. 2011. Perceived
predation risk reduces the number of offspring songbirds produce per year.
Science 334:1398–1401.

Associate Editor: DeStefano.

SUPPORTING INFORMATION

Additional supporting information may be found in the
online version of this article at the publisher’s web-site.

Step-by-step instructions for soldering the electrical con-
nection of a CD player play button that utilizes a tactile
momentary action pushbutton switch.

830 Wildlife Society Bulletin � 38(4)

https://climate.agry.purdue.edu/climate/narrative.asp
https://climate.agry.purdue.edu/climate/narrative.asp
https://climate.agry.purdue.edu/climate/summary.asp
https://climate.agry.purdue.edu/climate/summary.asp
http://www.ncdc.noaa.gov/oa/climate/normals/usnormals.html
http://www.ncdc.noaa.gov/oa/climate/normals/usnormals.html


R E S E A R C H A R T I C L E

Signatures of natural and unnatural selection: evidence
from an immune system gene in African buffalo

K. E. Lane-deGraaf • S. J. Amish • F. Gardipee •

A. Jolles • G. Luikart • V. O. Ezenwa

Received: 30 September 2013 / Accepted: 15 September 2014 / Published online: 16 October 2014

� Springer Science+Business Media Dordrecht 2014

Abstract Pathogens often have negative effects on

wildlife populations, and disease management strategies

are important for mitigating opportunities for pathogen

transmission. Bovine tuberculosis (Mycobacterium bovis;

BTB) is widespread among African buffalo (Syncerus

caffer) populations in southern Africa, and strategies for

managing this disease vary. In two high profile parks,

Kruger National Park (KNP) and Hluhluwe-iMfolozi Park

(HIP), BTB is either not actively managed (KNP) or

managed using a test-and-cull program (HIP). Exploiting

this variation in management tactics, we investigated

potential evolutionary consequences of BTB and BTB

management on buffalo by examining genetic diversity at

IFNG, a locus which codes for interferon gamma, a sig-

naling molecule vital in the immune response to BTB. Both

heterozygosity and allelic richness were significantly and

positively correlated with chromosomal distance from

IFNG in KNP, suggesting that directional selection is act-

ing on IFNG among buffalo in this park. While we did not

see the same reduction in genetic variation around IFNG in

HIP, we found evidence of a recent bottleneck, which

might have eroded this signature due to genome-wide

reductions in diversity. In KNP, alleles at IFNG were in

significant gametic disequilibrium at both short and long

chromosomal distances, but no statistically significant

gametic disequilibrium was associated with IFNG in HIP.

When, we compared genetic diversity between culled and

non-culled subsets of HIP animals, we also found that

individuals in the culled group had more rare alleles than

those in the non-culled group, and that these rare alleles

occurred at higher frequency. The observed excess of rare

alleles in culled buffalo and the patterns of gametic dis-

equilibrium in HIP suggest that management may be

eroding immunogenetic diversity, disrupting haplotype

associations in this population. Taken together, our results

suggest that both infectious diseases and disease manage-

ment strategies can influence host genetic diversity with

important evolutionary consequences.

Keywords Parasite-mediated selection � Syncerus caffer �
Bovine tuberculosis � IFNG � Disease management �
Genetic diversity � Gametic disequilibrium

Introduction

Parasites, including helminths, protozoa, bacteria, and

viruses, can act as strong selective forces on host

K. E. Lane-deGraaf and S. J. Amish have contributed equally to this

paper.

Electronic supplementary material The online version of this
article (doi:10.1007/s10592-014-0658-0) contains supplementary
material, which is available to authorized users.

K. E. Lane-deGraaf (&) � V. O. Ezenwa (&)
Odum School of Ecology and Department of Infectious

Diseases, College of Veterinary Medicine, University of

Georgia, Athens, GA 30602, USA

e-mail: lanedegraaf.kelly@gmail.com

V. O. Ezenwa

e-mail: vezenwa@uga.edu

S. J. Amish � G. Luikart
Fish & Wildlife Genomics Group, Division of Biological

Sciences, University of Montana, Missoula, MT 59812, USA

F. Gardipee

U.S. Fish and Wildlife Service, Sacramento, CA 59825, USA

A. Jolles

College of Veterinary Medicine and Department of Integrative

Biology, Oregon State University, Corvallis, OR 97331, USA

123

Conserv Genet (2015) 16:289–300

DOI 10.1007/s10592-014-0658-0

http://dx.doi.org/10.1007/s10592-014-0658-0


populations driving evolutionary change (Smith et al. 2009;

Koskella et al. 2012; Leung et al. 2012). Selection by

parasites may have particularly potent impacts on regions

of the genome that are directly involved in immune

defense. Immune system genes often show much greater

adaptive evolution than genes not directly involved in

immune function, and this pattern is often interpreted as a

signature of intense co-evolutionary interactions between

hosts and parasites (McTaggart et al. 2012). Parasite-

mediated selection on immune genes can occur directly

when parasites reduce individual survival or depress

fecundity, increasing the frequency of resistant genotypes

(Altizer et al. 2003; Thrall and Burdon 2003), and also

indirectly via management responses to infectious disease

outbreaks (Shim and Galvani 2009). In the latter case,

strategies used to control parasite spread, such as culling or

vaccination, may accelerate the rate and intensity of

observable selection or drive cryptic evolutionary change.

Parasites interact with the host immune system in dif-

ferent ways, resulting in distinct forms of selection acting

on immunity. Although immune system genes are expected

to experience strong selection as a group, the mode and

degree of selection can be highly variable across loci.

Balancing selection was once viewed as a dominant form

of selection acting on immune loci in wildlife populations

in particular, due in part to a historical focus of research on

major histocompatibility (MHC) genes (Acevedo-White-

house and Cunningham 2006). However, more recent work

on a broader set of immune function genes has expanded

this view (e.g. Downing et al. 2009; Tonteri et al. 2010;

Tschirren et al. 2011; Llewellyn et al. 2012). In the last five

years, more than one thousand studies have described

evidence of selection on immune genes in mammals

(excluding humans); of these studies, almost half found

evidence of directional selection, while less than a tenth

found evidence of balancing selection, and fewer still

found evidence of diversifying selection highlighting the

level of variability in the mode of selection acting on

immune genes. For example, in a study of 18 microsatel-

lites linked to immune genes in Atlantic salmon (Salmo

salar), Tonteri et al. (2010) found that genes for interleu-

kin1 and calmodulin production, were under strong direc-

tional selection. By contrast, a study of five different

immune genes in the greater prairie-chicken (Tympanachus

cupido), including interleukin 2 and transforming growth

factor b3, found evidence of directional selection at only
one of five genes (inhibitor of apoptosis protein-1), while

selection at the remaining loci was more consistent with

balancing selection (Bollmer et al. 2011). Finally, in an

analysis of the interleukin 4 (IL-4) locus, thought to be

under selective pressure from helminth parasites in mam-

mals, diversifying selection was found to play a role in

maintaining genetic diversity at fifteen IL-4 residues

involved in receptor binding. This pattern of diversifying

selection was maintained across multiple mammalian

genomes, ranging from mice to primates, carnivores and

ungulates (Koyanagi et al. 2010). These case studies sug-

gest that interactions between parasites and the host

immune system result not only in distinct signatures of

selection on different immune genes, but also that the form

of selection can vary across species and populations.

The environmental context of a particular host-parasite

interaction can play a key role in determining the strength

and form of selection acting on any immune gene. In recent

years, disease management has emerged as an important

factor potentially driving selection on immune loci (Altizer

et al. 2003; McCallum 2008). Management strategies for

controlling infectious diseases in natural populations are

becoming increasingly common, and as such, understand-

ing how management can shape the evolution of host

immune defenses is critical (Woodroffe 1999; Cross et al.

2009; Joseph et al. 2013). Culling, in particular, is one

disease management strategy used in wildlife that has

enormous potential to drive selection on hosts, particularly

when specific groups of individuals are targeted for

removal (Myerstud and Bischof 2010; Myerstud 2011;

White et al. 2011). While a number of studies have docu-

mented unintended ecological effects of culling (Donnelly

et al. 2006; Woodroffe et al. 2006, 2009a, b), empirical

evidence of evolutionary consequences are still very lim-

ited (Smith et al. 2009). However, culling has the potential

to erode host evolutionary potential by facilitating the loss

of genetic diversity via reductions in rare alleles (Sackett

et al. 2013), impeding the evolution of resistance (Shim

and Galvani 2009), and even driving broad evolutionary

changes in the ecological community if disease manage-

ment for one species changes the demography and ecology

of non-target species (Chauvenet et al. 2011).

In this study, we focus on bovine tuberculosis (BTB) as

a potential agent of selection on immune genes in a wild

reservoir host population. BTB is a chronic disease of

wildlife and livestock, caused by Mycobacterium bovis.

The disease has a global distribution and accounts for

significant economic losses worldwide (Michel et al. 2010;

Goodchild et al. 2011). In South Africa, African buffalo

(Syncerus caffer) serve as the primary wildlife reservoir of

BTB and are responsible for spillover into other wildlife

populations (e.g. lions, cheetahs) and cattle (Renwick et al.

2007; Fitzgerald and Kaneene 2013; de Garine-Wichatit-

sky et al. 2013). As such, effective disease control in

buffalo populations is of critical importance (de Garine-

Wichatitsky et al. 2013). Management strategies for con-

trolling BTB in South African buffalo range from passive

surveillance to active test and cull programs (Michel et al.

2006). For instance, in Kruger National Park (KNP), where

BTB was first detected in the buffalo population between

290 Conserv Genet (2015) 16:289–300

123



1950 and 1960 and the population is estimated to be

between 23,000 and 25,000 individuals, BTB management

has focused on surveillance, population monitoring, and

research (Michel et al. 2006). By contrast, in Hluhluwe-

iMfolozi Park (HIP) where the first BTB positive buffalo

was detected in 1986 and the buffalo population is con-

siderably smaller (est. 3,000 individuals), a test and cull

program was initiated in the mid-1990s (Michel et al.

2006). Current estimates of BTB prevalence in both parks

are between 5 and 45 % for buffalo herds in KNP (Cross

et al. 2009), and 0–73 % for buffalo herds in HIP (Jolles

et al. 2006). Since negative effects of BTB on survival and

reproduction have been described for buffalo (Jolles et al.

2005), it is possible that this disease could act as a direct

selective force on buffalo populations. Moreover, in HIP

where the BTB control program culled approximately 700

buffalo testing positive for BTB (out of a total of 4,681

animals tested between 1999 and 2006; Jolles et al.,

unpublished data), selective culling could impose addi-

tional indirect selective pressure, particularly at loci

involved in immune defense against the disease. Thus, the

African buffalo-BTB system presents an ideal platform for

studying potential direct and indirect effects of parasites on

the evolution of immune genes in the wild.

Taking advantage of this unique study system, we tested

for evidence of selection on the interferon gamma (IFNG)

gene in populations of African buffalo in KNP and HIP.

The IFNG locus codes for an immune signaling molecule

of the same name that plays an important role in the

response to M. bovis infection. IFNc is a key cytokine
involved in T helper (TH1) cell responsiveness that is

triggered by intracellular pathogens, including M. bovis

(Waters et al. 2012). Given the critical role of IFNc in
pathogen defense generally, and the response to BTB

specifically, variability in the IFNc phenotype is likely to
be associated with fitness in the face of infection. Impor-

tantly, BTB diagnosis in buffalo relies on IFNc-based tests
that measure the strength of an individual’s immune

response to M. bovis antigen challenge (Wood et al. 1991;

Ryan et al. 2000; Cousins and Florisson 2005), thus culling

based on variation in this response could impose strong

selection on the IFNG locus.

To explore the possibility that BTB, BTB-related cull-

ing, or both could be driving selection at IFNG, and to

identify the form of selection that might be acting, we

quantified genetic diversity (heterozygosity and allelic

richness) and gametic (linkage) disequilibrium (GD) across

neutral loci flanking IFNG to examine how diversity and

GD change with distance around a locus putatively under

selection. First, we investigated patterns of genetic diver-

sity and GD around IFNG in the total population at both

parks to evaluate how intracellular pathogens in general

(and BTB specifically) may be driving selection at immune

genes, irrespective of management strategy. Second, we

explored whether culling had additional effects on IFNG

by examining whether the number and frequency of rare

alleles at IFNG and surrounding loci were directly

impacted by culling. We predicted that parasite-mediated

selection would result in a distinct signature of selection at

IFNG and nearby loci in both parks. Specifically, we

expected that if balancing selection is occurring, we would

see higher levels of diversity at IFNG relative to sur-

rounding loci. By contrast, if directional selection is

occurring, we would see lower levels of diversity at IFNG

compared to flanking loci. With respect to GD, we

expected that directional selection might result in stronger

patterns of gametic disequilibrium involving IFNG versus

flanking loci. Finally, we also predicted that if disease

management contributes to selection at IFNG this would be

evident as a stronger selection signature in the HIP popu-

lation where culling takes place. Alternatively, culling

might produce relatively cryptic genetic changes in this

population, disrupting patterns genetic diversity and GD

around IFNG.

Methods

Sample collection

We sampled buffalo at two sites, Hluhluwe-iMfolozi Park

(HIP) and Kruger National Park (KNP) in South Africa. In

HIP, males and females were captured in the Masinda

section of the park as part a Bovine Tuberculosis Control

Program in 2005 and 2006. In KNP, female buffalo were

captured in the Lower Sabie and Crocodile Bridge regions

in 2008 as part of a research study. Captures in HIP were

carried out by park management using a helicopter and

funnel system to drive herds into a capture corral. In KNP,

animals were darted from a helicopter by the South Africa

National Parks Veterinary Wildlife Services. Whole blood

from HIP buffalo (n = 83) was preserved on FTA cards

(Whatman
�

Inc, Clifton, NJ, USA), and dried cards were

stored at room temperature for one year until DNA

extraction. In KNP, tissue samples were collected from

buffalo (n = 209) and stored in 2 ml tubes with silica gel

at room temperature for up to 24 months prior to DNA

extraction.

In HIP, all captured buffalo were tested for bovine

tuberculosis using a tuberculin skin test. Briefly, individual

buffalo were injected with bovine tuberculin intra-der-

mally, and a localized swelling response measured 72 h

later (Ryan et al. 2000). Animals were considered

BTB ? if the swelling response was greater than 2 mm.

Skin test-positive buffalo were culled as part of the BTB

control program.

Conserv Genet (2015) 16:289–300 291

123



Molecular methods

We focused on the IFNG gene which codes for IFNc, a
protein critical in the immune response to a variety of

intracellular pathogens, including M. bovis (Bradley et al.

1996, Bream et al. 2000, Pollock et al. 2008). Twelve

flanking microsatellite loci were chosen based on their

proximity to IFNG, and range in distance from 3.1

(BMS1617) to 28.4 (KERA) cM on either side of IFNG

(See Table S1). Studies of cattle and sheep suggest that all

but one of these 12 flanking loci are neutral, or functionally

neutral, genes. BL4 has been associated with immunity and

disease resistance in sheep and African buffalo (Coltman

et al. 1999, 2001, Ezenwa et al. 2010); while the following

six loci have been reported to be neutral: BMS1617

(Kappes et al. 1997), RM154 (Maddox et al. 2001),

BR2936 (Kappes et al. 1997), KRT2 (Maddox et al. 2001),

ILSTS22 (Kappes et al. 1997), AGLA293 (Kappes et al.

1997). We considered five additional loci to be functionally

neutral based on their involvement or proximity to genes

with functions not related to immunity or disease resis-

tance, including: CSSM34, which codes for a retinoic acid

receptor important in Vitamin A absorption (Barendse

2002); GLYCAM1, KERA, and TEX15, which code for

proteins important in lactation, corneal development, and

chromosomal synapsis and meiotic recombination,

respectively (Groenen et al. 1995; Tocyap et al. 2006;

Yang et al. 2008); and IGF, which codes for insulin-growth

factor, which modulates cell growth and is linked with

increased tumor development (Renehan et al. 2004).

DNA was extracted from tissue samples using the QIA-

GEN Blood & Tissue Kit (Valencia, CA) following the

manufacturer’s protocol; for FTA cards, a modified protocol

was used (Lisette Waits, pers. comm). Multiplex PCRs were

optimized, and 10ul reactions were performed in MJR

PTC200 thermocyclers. Each reaction contained: 1ul of

template DNA, 4.5 ul of QIA multiplex mix (Qiagen), and 1 ul

of 2 pM forward and reverse primers. Two different touch-

down profiles with 35–40 cycles were used, one with an initial

annealing temperature of 64 �C stepping down to 59 �C, and
another starting at 58 �C and stepping down to 53 �C. Fluo-
rescently-labeled DNA fragments were visualized on an

ABI3130xl automated capillary sequencer (Applied Biosys-

tems). Allele sizes were determined using the ABI GS600LIZ

ladder (Applied Biosystems). Chromatograms were analyzed

and confirmed by two independent technicians using

GeneMapper software v3.7 (Applied Biosystems).

Locus descriptions

We tested for the presence of null alleles at all loci using

Micro-Checker v. 2.2.3 (van Oosterhout et al. 2004), and

for deviations from Hardy–Weinberg proportions (HWP)

using GENEPOP v. 4.1 (Raymond and Rousset 1995). No

null alleles were found at any locus. All loci were in HWP

except for AGLA293 (p \ 0.001 in both populations;
Table 1), which had a heterozygote deficit, so further

analyses were run both with and without this locus. Since

inclusion of AGLA293 did not qualitatively change our

results, we only report the results including all loci. We

quantified the magnitude of deviation from HWP by

computing FIS in each population. We also calculated FST
values to evaluate the degree of genetic differentiation

between the two study populations. Both FIS and FST val-

ues were calculated using Fstat v. 2.9.3 (Goudet 2001).

Finally, mean population relatedness was calculated in

Genalex using the Lynch & Ritland estimator multiplied by

two so that it scales from zero to one with full sibs sharing

half their alleles (r = 0.5).

Indices of genetic diversity and GD

To evaluate whether selection has affected the IFNG locus,

we calculated two measures of genetic diversity for IFNG

and the 13 flanking loci. First, we calculated allelic richness

(AR) at each locus using Fstat v. 2.9.3 (Goudet 2001). Next,

we estimated the heterozygosity at each locus by calculating

both observed and expected heterozygosity under Hardy–

Weinberg proportions. Observed and expected heterozy-

gosities (HO and HE, respectively) were calculated in Arle-

quin v. 3.1 (Schneider et al. 2000), but subsequent analyses

are limited to HE as it more accurately reflects genetic vari-

ation within the population as a whole (Nei 1987).

We calculated pairwise gametic disequilibrium, a mea-

sure of non-independence between loci due to either prox-

imity or function. GD was expressed as D0, a derivative of D
which is the deviation in the frequency of co-occurrence

between multiple alleles (i.e. haplotypes) due to gametic

disequilibrium (Lewontin and Kojima 1960). D0 is defined
as D0 = D/Dmax where Dmax is the maximum value of D,
given a set of allele frequencies, and D0 = 1 is complete
gametic disequilibrium (Lewontin 1964). In addition to D0,
we also calculated GD as r

2
, which is a measure of GD that

is influenced by allele frequencies. However, results for r
2

were not quantitatively distinct from D0, and as such, we
report only the results of D0. D0 and r2 values were calcu-
lated using PowerMarker (Liu and Muse 2005). The sig-

nificance of pairwise D0 estimates were evaluated using
Genepop (v4.2), and accepted at p B 0.0002.

Patterns of selection

To test for a signature of selection within each study popu-

lation, we examined the association between genetic diver-

sity at each locus and its chromosomal distance (the absolute

value in cM) from the putative gene under selection (IFNG).

292 Conserv Genet (2015) 16:289–300

123



This use of multiple loci across the gene region helps control

for genome-wide (e.g. demographic) variations and has been

previously used to identify loci under selection and selective

sweeps (Ihle et al. 2006; Makinen et al. 2008). We tested for

correlations between distance and allelic richness and het-

erozygosity using Spearman rank tests. Since differences in

chromosomal distance between loci should reflect differ-

ences in the rate of recombination and/or selection, we also

tested for associations between statistically significant val-

ues of pairwise D0 and the distance (cM) between locus pairs.
Linear regression tests were used to evaluate whether levels

of GD decayed with increasing distance between loci.

Genetic diversity in HIP

We compared genetic diversity between the two study pop-

ulations using Wilcoxon paired sign rank tests. In HIP, where

overall genetic diversity was much reduced, we tested for

evidence of potential bottlenecks using heterozygosity excess

and deficiency tests in Bottleneck (v. 1.2). We used a two-

phase model of microsatellite evolution, a biologically

appropriate model that captures the evolution of microsatel-

lites (Cornuet and Luikart 1996; Luikart and Cornuet 1998).

We parameterized the model by defining 80 % of mutations

as conforming to a stepwise mutation model (Kimura and

Ohta 1978) and 20 % to a multistep model, assuming a var-

iance of 12 for the geometric distribution of number of repeat

units per multi-step mutation. Mode shift tests were used to

assess bottleneck strength (Cornuet and Luikart 1996).

To identify potential effects of culling on genetic

diversity in HIP, we compared our measures of diversity

(allelic richness and heterozygosity) between two subsets

of the HIP population: animals that were culled as a result

of a positive tuberculin skin test (BTB positive, n = 11),

and animals that were poor reactors on the skin test and

were not culled (BTB negative, n = 64). Comparisons

between population subsets were done using Wilcoxon

paired signed rank tests. We also examined the occurrence

of rare alleles in the culled and non-culled population

subsets of HIP to test if rare alleles are being removed

disproportionately as a result of culling, potentially con-

tributing to an overall erosion of genetic diversity in the

park. To do this, for each population subset, we calculated

the number and frequency of alleles at each locus with a

frequency of less than 0.1. We then tested for locus-specific

differences in the number and frequency of these rare

alleles in the two population subsets using Wilcoxon paired

signed rank tests.

Results

Locus descriptions

Thirteen microsatellite loci spanning IFNG were geno-

typed successfully in a total of 292 individuals from two

populations (KNP: n = 209; HIP: n = 83). Locus-specific

FIS ranged from -0.155 to 0.357 in HIP and from -0.102

to 0.183 in KNP (Table 1). Low and/or negative FIS values

found at 9 of 13 loci in HIP and 6 of 13 loci in KNP

indicate that there is little to no cryptic subpopulation

structuring occurring within these populations. Pairwise

Table 1 Population structure statistics for HIP and KNP

Locus cM from IFNG HWP FIS FST AR HO HE

HIP KNP HIP KNP HIP KNP HIP KNP HIP KNP

TEX15 -26 0.19 0.06 0.057 0.023 0.048* 5 8.97 0.638 0.796 0.675 0.815

IGF -19.4 0.99 0.91 0.023 0.052 0.149* 3 6 0.627 0.711 0.641 0.750

RM154 -15.6 0.70 0.12 -0.029 0.029 0.115* 7.81 15.36 0.795 0.871 0.773 0.897

BR2936 -7.1 0.24 0.56 -0.090 0.027 0.071* 5 7.95 0.866 0.731 0.795 0.752

BMS1617 -3.1 0.29 0.21 -0.155 -0.102 0.010 2 2 0.361 0.253 0.313 0.229

IFNG 0 1.00 0.08 -0.043 -0.071 0.056* 2 4.72 0.096 0.368 0.091 0.344

BL4 ?3.6 0.32 0.40 0.097 -0.021 0.089* 7 8.59 0.676 0.836 0.749 0.819

CSSM34 ?10.9 0.79 0.99 -0.038 0.037 0.050* 4.99 8.02 0.687 0.699 0.662 0.716

KRT2 ?15.2 0.59 0.76 -0.056 -0.015 0.080* 4.82 8.20 0.771 0.754 0.730 0.743

ILSTS22 ?19.1 0.48 0.31 -0.056 -0.013 0.032* 3 7.47 0.542 0.638 0.514 0.632

GLYCAM1 ?19.7 0.74 0.56 -0.085 -0.046 0.095* 4.97 10.90 0.795 0.909 0.733 0.869

AGLA293 ?21.3 0.00* 0.00* 0.357 0.183 0.180* 6.66 13.78 0.415 0.701 0.643 0.868

KERA ?28.4 0.41 0.41 -0.033 0.011 0.062* 7.98 11.57 0.867 0.839 0.839 0.848

? and - signs denote chromosomal distance from IFNG in centimorgans (cM). * denote significant (p B 0.05) deviations from Hardy–Weinberg

proportions (HWP) and significant genetic structure (pairwise FST). Also reported are FIS values, allelic richness (AR), observed and expected

heterozygosity (HO, HE)

Conserv Genet (2015) 16:289–300 293

123



FST values between populations ranged from 0.010 to

0.1804 among loci (Table 1). There was evidence of sig-

nificant genetic differentiation between the two populations

at all but one locus; BMS1617 was the only locus with a

non-significant FST value (FST = 0.010; p = 0.075). This

locus also had the lowest FIS value in both populations

(FIS KNP = -0.155; FIS HIP = -0.102; Table 1). Mean

relatedness among individuals in HIP was 0.185 and 0.019

in KNP.

Signature of selection: patterns of genetic diversity

and GD

Overall, genetic diversity was lower at IFNG compared to

other loci in both populations. AR ranged from 2 to 7.97

alleles per locus in HIP and from 2 to 15.36 in KNP

(Table 1). Excluding BMS1617 for which we found no

significant evidence of genetic differentiation between the

two parks, IFNG was the locus with the lowest AR value in

both populations, with only 2 alleles identified in HIP and

4.7 alleles in KNP. Similarly, expected heterozygosity was

also lowest at IFNG in both populations (0.091 in HIP,

0.344 in KNP; Table 1). Observed heterozygosity showed

similar patterns as expected heterozygosity.

By examining genetic variation across loci at increasing

distances from IFNG, we found evidence of a signature of

selection in one population but not the other. In KNP, both

allelic richness and heterozygosity were significantly and

positively correlated with distance from IFNG (Spearman

rank correlation: AR: rho = 0.637, p = 0.019; HE:

rho = 0.593, p = 0.032; Fig. 1a–b). Although the distance

pattern observed for allelic richness and heterozygosity in

HIP mirrored the pattern observed in KNP, no significant

associations were detected between either measure of

genetic diversity and distance from IFNG in the HIP pop-

ulation (AR: rho = 0.463, p = 0.111; HE: rho = 0.324,

p = 0.279; Fig. 1 a–b). Given the reduced genetic diversity

at BMS1617 relative to IFNG and other surrounding loci in

KNP, we also examined whether the patterns we observed

for IFNG could have been driven by BMS1617. To do this

we re-ran the distance analyses (e.g., association tests) using

BMS1617 as the target locus. We found no evidence of an

association between chromosomal distance from BMS1617

and genetic diversity (AR: rho = 0.545, p = 0.066; HE:

rho = 0.486, p = 0.106). This supports the supposition that

IFNG, and not BMS1617, is the putative target of selection

in the region under analysis.

Fifteen pairs of loci were in significant GD in HIP, with

the average D0 = 0.63; by contrast, seven pairs of loci were
in significant GD in KNP, with the average D0 = 0.44
(Table 2). The two populations shared only four of 22

significant locus pairs (Table 2), with two of these showing

similar deviations in haplotype frequency across

populations (ILSTS22-AGLA293 and KRT2-AGLA293),

and two showing higher levels of D0 in HIP (ILSTS22-
GLYCAM1 and RM15-BR2936; Table 2). The median

distance between loci with significant pairwise GD was

5.8 cM in HIP and 7.3 cM in KNP, with pairwise distances

ranging from 0.6 to 19.7 cM (Table 2). In KNP, the

strongest deviation in haplotype frequencies was between

IFNG and BL4 (D0 = 0.572; 3.6 cM), while the longest
significant GD observed was between IFNG and GLY-

CAM1 (19.7 cM). In HIP, the strongest deviation in hap-

lotype frequencies was between ILSTS22 and GLYCAM1

(D0 = 0.987; 0.6 cM), and there was no significant pair-
wise GD at distances greater than 13.2 cM. When we

tested for associations between GD and the chromosomal

distance between loci, we found that GD declined signifi-

cantly with pairwise distance in KNP (n = 7, r
2

= 0.77,

p = 0.0094; Fig. 2), but not HIP (n = 15, r
2

= 0.18,

p = 0.1183; Fig. 2).

Reduced genetic diversity in HIP

Two tests—the mode shift and heterozygosity-excess

tests—revealed evidence of a recent bottleneck in HIP

(mode shift: bimodal distribution; heterozygosity excess

Fig. 1 Patterns of a allelic richness and b expected heterozygosity at
increasing chromosomal distance (cM) from IFNG in KNP and HIP.

HIP is represented by the dashed line, and KNP is represented by the

solid line

294 Conserv Genet (2015) 16:289–300

123



test: p \ 0.05; see Table S2). Furthermore, a comparison of
genetic diversity between the HIP and KNP populations

indicated that diversity is reduced in HIP compared to

KNP. Both allelic richness and expected heterozygosity

were significantly lower in HIP compared to KNP (Wil-

coxon paired signed rank test: AR, S = 39.0, p = 0.0005;

HE, S = 36.5, p = 0.0081).

While the bottleneck in HIP may account for the pattern

of eroded genetic diversity in this population, BTB man-

agement (i.e. culling) could also contribute to the overall

reduction in genetic diversity. To explore this possibility,

we tested for differences in genetic diversity between

culled (C) and non-culled (NC) groups of individuals from

HIP. Although there was no difference between population

subsets in either diversity index (Wilcoxon signed rank

test: AR: S = -18.0, p = 0.229; HE: S = -5.0,

p = 0.734), the culled group had no heterozygosity at

IFNG, possibly indicating strong selection acting on this

locus.

Focusing on rare alleles, we found that individuals in the

culled group had a significantly higher number of rare

alleles at loci surrounding IFNG (C = 26; NC = 22;

S = 13.5, p = 0.0358; Table 3). Moreover, for those rare

alleles present in both the culled and non-culled groups,

over 70 % (10 out of 14) occurred at a higher frequency in

the culled group (Table 3). Rarity is influenced by the

number of alleles per locus and therefore could vary

between population subsets simply because of the greater

number of alleles available to sample in larger populations.

However, even when we considered only the seven rare

alleles that were present in both population sub-groups, and

that occurred at loci where all alleles were shared between

groups, over 80 % (5 out of 6) occurred at a higher fre-

quency in the culled group (Table 3), suggesting that rare

alleles were overrepresented in the culled subset of the

population and that culling may be disproportionately

eliminating these alleles.

Discussion

Our findings suggest that directional selection is acting on

and around the IFNG locus in the buffalo population of

Kruger National Park. In the Hluhluwe-iMfolozi Park

population, reduced genetic diversity due to recent bottle-

neck events may have masked any signature of directional

selection driven by disease. However, we found evidence

that disease management may be compounding the loss of

genetic diversity in the IFNG gene region. In particular,

culling to reduce bovine tuberculosis (BTB) may be

selectively removing rare alleles from the HIP buffalo

population, prolonging genetic recovery from a recent

reduction in population size. Overall, our results suggest

that disease can directly or indirectly drive selection at

immune loci in wild populations, and that disease man-

agement might result in unintended evolutionary

Table 2 Locus pairs with significant levels of gametic disequilibrium
(GD) in HIP and KNP

Population Locus 1 Locus 2 CM D0

HIP ILSTS22 GLYCAM1 0.6 0.987

GLYCAM1 AGLA293 1.6 0.634

ILSTS22 AGLA293 2.2 0.418

KRT2 ILSTS22 3.9 0.707

BR2936 BMS1617 4 0.800

CSSM34 KRT2 4.3 0.727

KRT2 GLYCAM1 4.5 0.655

KRT2 AGLA293 6.1 0.456

TEX15 IGF 6.6 0.504

AGLA293 KERA 6.9 0.670

BL4 CSSM34 7.3 0.435

RM154 BR2936 8.5 0.668

GLYCAM1 KERA 8.7 0.587

ILSTS22 KERA 9.3 0.620

KRT2 KERA 13.2 0.528

KNP ILSTS22 GLYCAM1 0.6 0.556

ILSTS22 AGLA293 2.2 0.434

IFNG BL4 3.6 0.572

KRT2 AGLA293 6.1 0.437

RM154 BR2936 8.5 0.428

CSSM34 AGLA293 10.4 0.359

IFNG GLYCAM1 19.7 0.276

Chromosomal distance between loci is listed in centimorgans (cM)

and the strength of GD is measured as D0. The shortest distance
between possible pairwise comparisons was 0.6 cM while the longest

distance was 54.4 cM

Fig. 2 Relationship between pairwise gametic disequilibrium (GD)
and chromosomal distance (cM) in KNP and HIP. HIP is represented

by the dashed line with open circles, and KNP is represented by the

solid line with filled circles

Conserv Genet (2015) 16:289–300 295

123



consequences by exacerbating underlying population

demographic effects.

Evidence of selection in KNP

Our conclusion that selection is acting on IFNG in buffalo

comes from several lines of evidence. By examining 12

loci flanking IFNG, we uncovered a significant, positive

relationship between chromosomal distance from IFNG

and both allelic richness and heterozygosity in the KNP

population. The positive relationship between genetic

diversity and chromosomal distance from IFNG suggests

that directional selection (e.g. a selective sweep) is

occurring at this locus. In addition, we observed significant

gametic disequilibrium (GD) between IFNG and BL4

which is likely the effect of directional selection at IFNG

and genetic hitchhiking at BL4. While there was a signif-

icant association between pairwise chromosomal distance

between loci and GD in KNP, the level of GD between

IFNG and BL4 was higher than for any other pair of loci,

irrespective of distance. This suggests that in this region of

the chromosome, selection may be generating a non-ran-

dom association between alleles at these two loci which is

stronger than that expected based on distance alone.

Interestingly, despite this signature of directional selection,

several low frequency alleles remain at IFNG in the KNP

population (three alleles at \2 % frequency), and the long
distance GD observed between IFNG and GLYCAM1

(19.7 cM apart) suggests that haplotypes involving these

low frequency alleles are being maintained in KNP.

In contrast to the genetic diversity-chromosomal dis-

tance pattern we observed in KNP, no such pattern was

evident in HIP. However, an overall reduction in genetic

diversity in the IFNG gene region in the HIP population

could have masked any distance effect. In the mid-1950s

the HIP buffalo population dropped to as few as 800

individuals followed by rapid population growth thereafter

(Jolles 2007); this event and the elevated levels of relat-

edness among HIP individuals corroborates our evidence

for a recent bottleneck in the population. Bottlenecks and

founder events can drastically reduce the amount of genetic

diversity in populations (Maruyama and Fuerst 1984),

Table 3 Number and
frequency of rare alleles in HIP,

parsed by culled (C, n = 11)

and non-culled (NC, n = 64)

population segments

Zeros represent alleles that are

absent from the associated

population segment

S shared rare alleles

Loci NA Rare NA Rare alleles Frequency

Total NC C Total NC C T NC C

IFNG 2 2 1 1 1 0 IFNG allele1 0.045 0.052 0

BMS1617 2 2 2 0 0 0 NA

BL4 7 7 7 3 3 4 BL4 allele1 0.032 0.027
S

0.063
S

BL4 allele2 0.063 0.064
S

0.063
S

BL4 allele3 0.024 0.018
S

0.063
S

BR2936 5 5 5 0 0 1 NA

CSSM34 5 5 4 2 2 1 CSSM allele1 0.086 0.077 0.136

CSSM allele2 0.026 0.031 0

KRT2 5 5 4 1 1 1 KRT2 allele1 0.007 0.008 0

RM154 8 8 6 5 5 6 RM154 allele1 0.086 0.085
S

0.091
S

RM154 allele2 0.086 0.092
S

0.045
S

RM154 allele3 0.007 0.008 0

RM154 allele4 0.059 0.054
S

0.091
S

RM154 allele5 0.013 0.015 0

ILSTS22 3 3 3 1 0 1 ILSTS allele1 0.099 0.1 0.091

IGF 3 3 3 0 0 1 NA

GLYCAM1 5 5 4 1 1 2 GLYCAM allele1 0.013 0.015 0

AGLA293 7 7 5 4 4 4 AGLA allele1 0.093 0.094
S

0.091
S

AGLA allele2 0.007 0.008 0

AGLA allele3 0.02 0.016
S

0.045
S

AGLA allele4 0.007 0.008 0

TEX15 5 5 5 2 2 2 TEX allele1 0.075 0.073 0.091

TEX allele2 0.055 0.04 0.136

KERA 8 8 7 3 3 3 KERA allele1 0.086 0.069 0.182

KERA allele2 0.02 0.015
S

0.045
S

KERA allele3 0.007 0.008 0

296 Conserv Genet (2015) 16:289–300

123



making signatures of selection difficult to detect (Frere

et al. 2011). Nevertheless, characteristics of individual loci

hint that selection on IFNG could be occurring in HIP, as in

KNP. Specifically, in both populations, heterozygosity at

IFNG was at or close to the lowest values recorded across

all loci (KNP = 0.344, range: 0.229–0.897; HIP = 0.091,

range: 0.091–0.839, Table 1), a pattern suggestive of

directional selection acting on IFNG in both populations.

Evidence of directional selection acting on immune loci

in response to bacterial pathogens, including M. bovis, has

been reported in livestock, suggesting that disease, possibly

BTB, could be driving the pattern of selection we observed

at IFNG. For example, BTB has been implicated as a

potential force driving directional selection of the disease

resistance gene, NRAMP1, in African Zebu cattle (Kad-

armideen et al. 2011). Evidence of directional selection has

also been found at the porcine TLR-4 gene in response to

gram-negative bacterial pathogens (Palermo et al. 2009).

Thus, it is possible that BTB infection acts to reduce

genetic diversity at the IFNG locus in buffalo, resulting in

the conservation of only those alleles important in

mounting an effective immune response, as has been

shown in for the PRNP gene in cervids in North America in

response to chronic wasting disease (Robinson et al. 2012).

Management-driven selection in HIP

Effects of disease management on genetic variation are

poorly understood, rarely assessed, but potentially strong

(Allendorf and Hard 2009). In HIP, we exploited a rare

opportunity to assess the role of disease management in

driving indirect selection. Although there was no clear

signature of selection in this population, we did find evi-

dence that disease management (i.e. culling) might affect

diversity at immune system genes. Specifically, we found

that although there were no differences in allelic richness

or heterozygosity between culled and non-culled segments

of the HIP population, the culled segment had a signifi-

cantly greater number and frequency of rare alleles sug-

gesting that these alleles are being lost from the population

via culling. A comparison of GD patterns in the two pop-

ulations further suggests that culling may have disrupted

the pairwise distance-GD relationship in HIP by eliminat-

ing haplotypes that are being maintained between IFNG

and surrounding loci (e.g. GLYCAM1) in KNP.

Evidence that IFNG haplotypes being maintained in

KNP have been lost from HIP suggests that the rare allele

loss observed in HIP may have important fitness conse-

quences. Rare allele advantage can be an important com-

ponent of a population’s ability to respond to infectious

disease agents (Spurgin and Richardson 2010; Lankau and

Strauss 2010). Thus, the reduction in rare alleles in HIP

could have implications for population level responses to

future infectious disease threats. Given the small sample

size we used for the rare alleles comparison (11 vs. 64), we

recognize that these results need to be interpreted with

caution. Nevertheless, based on the strong preliminary

patterns we report here and in light of work showing that

rare alleles can confer fitness advantages, particularly with

respect to response to pathogens (Koskella and Lively

2009; Sommer 2005), the potential loss of rare alleles due

to culling in the buffalo-BTB system deserves further

investigation.

The bottleneck in the mid-1950s, and subsequent rapid

growth of the HIP population, likely resulted in the founder

event and reduction in overall genetic diversity that we

observed in the park. Culling could be prolonging this

bottleneck event by eliminating rare alleles from the pop-

ulation and reducing heterozygosity at IFNG. Disease

management in this population has had effects on the

population ecology of buffalo, with rapid population

declines and reduced population growth rates typically

following culling events (Jolles 2007). In addition to these

ecological effects, our results suggest that culling is also

affecting buffalo population genetics and long-term evo-

lutionary potential. Management strategies for controlling

invasive and emerging diseases in wildlife will continue to

be important due to the risk of disease spillover into live-

stock, wildlife, and humans (Michel et al. 2006; Smith

et al. 2009). For buffalo in HIP, however, culling for dis-

ease management may also be sustaining the effects of a

historical population bottleneck. More generally, our find-

ings suggest that there is real potential for unforeseen

evolutionary consequences to arise from disease manage-

ment in wildlife populations.

There are several notable examples of unintended eco-

logical consequences of culling (Singleton et al. 2007;

Bowen 2013). One well-known example is badger culling

in the UK, which was intended to limit spread of BTB to

cattle, but was instead linked to increases in transmission

(Donnelly et al. 2003, 2006; Pope et al. 2007; Woodroffe

et al. 2006, 2009a, b). Far less is known about the evolu-

tionary consequence of culling as a disease management

strategy. Sport hunting, which in some instances has the

same characteristics as culling, has been shown to have a

variety of unintended outcomes, including altering gene

flow between populations and, most importantly, selec-

tively removing individuals with desired traits (Harris et al.

2002; Allendorf and Hard 2009). For example, hunter

selection, in conjunction with deteriorating environmental

conditions, led to a significant reduction in horn size

among bighorn sheep rams in Arizona (Hedrick 2011).

Because BTB-culling programs that use IFNc-based diag-
nostic tests will selectively remove individuals that pro-

duce high levels of IFNc in response to an antigen
challenge (Wood et al. 1991; Ryan et al. 2000; Cousins and

Conserv Genet (2015) 16:289–300 297

123



Florisson 2005), an analogous potential consequence of

BTB-culling could be the selective removal of individuals

capable of mounting strong immune responses to BTB

infection. This selection could lead to significant popula-

tion biases in immune profiles. Future research is needed to

examine whether genetic differences in culled versus non-

culled buffalo translate into observable phenotypic differ-

ences in immune responses to BTB.

Conclusions

Opportunities for disease transmission between wildlife,

livestock, and humans are growing. However, the effects of

disease and disease management on evolution in wildlife

populations remain poorly understood. In particular, strat-

egies used to control disease might unwittingly lead to

cryptic evolutionary change (Shim and Galvani 2009). We

found that in an African buffalo population where BTB has

been present for over 50 years, and where there is no active

disease management, there was reduced diversity at (and

near) IFNG, a locus involved in mounting an immune

response to pathogens like tuberculosis. By contrast, in a

population where disease-based culling occurs, reduced

genetic diversity may be masking a signature of selection.

Furthermore, preliminary patterns show that management

actions, and thus unnatural selection, could be removing

rare alleles from the population, potentially driving cryptic

evolutionary change. While disease is widely recognized as

a potentially powerful force of selection, our results sug-

gest that disease management strategies can also have

important evolutionary consequences.

Acknowledgments We thank Kwazulu-Natal Wildlife Service,
Kwazulu-Natal State Veterinary Service, and South Africa National

Parks Veterinary Services for facilitating this research. In particular,

D. Cooper, P. Buss, and M. Hofmeyr provided invaluable assistance.

We thank Matt Gruber for his assistance in the lab. Animal protocols

used in this study were approved by the University of Georgia, the

University of Montana and Oregon State University Animal Care and

Use Committees. This work was supported by the National Science

Foundation Ecology of Infectious Diseases Program (DEB-1102493/

EF-0723928, EF-0723918), an NSF Small Grant for Exploratory

Research (DEB-0541762, DEB-0541981) and a University of Mon-

tana Faculty Research Grant.

References

Acevedo-Whitehouse K, Cunningham A (2006) Is MHC enough for

understanding wildlife immunogenetics? TREE 21(8):433–438

Allendorf FW, Hard JJ (2009) Human-induced evolution caused by

unnatural selection through harvest of wild animals. PNAS

106:9987–9994

Altizer S, Harvell D, Freidle E (2003) Rapid evolutionary dynamics

and disease threats to biodiversity. TREE 18(11):589–596

Barendse W (2002) Development and commercialization of a genetic

marker for marbling of beef in cattle: a case study. Intellectual

property rights animal breeding and genetics., pp 197–212

Bollmer JL, Ruder EA, Johnson JA, Eimes JA, Dunn PO (2011) Drift

and selection influence geographic variation at immune loci of

prairie-chickens. Mol Ecol 20:4695–4706

Bowen WD (2013) Marine mammal culling programs: review of

effects on predator and prey populations. Mamm Rev 43:207–

220

Bradley LM, Dalton DK, Croft M (1996) A direct role of IFN-gamma

in regulation of TH1 cell development. J Immunol 157(4):

1350–1358

Bream JH, Carrington M, O’Toole S, Dean M, Gerrard B, Shin HD,

Kosack D, Modi W, Young HA, Smith MW (2000) Polymor-

phisms of the human IFNG gene noncoding regions. Immuno-

genetics 51(1):50–58

Chauvenet ALM, Durant SM, Hilborn R, Pettorelli N (2011)

Unintended consequences of conservation actions: managing

diseases in complex ecosystems. PLoS One. doi:10.1371/journal.

pone.0028671

Coltman DW, Bowen WD, Wright JM (1999) A multivariate analysis

of phenotype and paternity in male harbor seals, Phoca vitulina,

at Sable Island, Nova Scotia. Behav Ecol 10:169–177

Coltman DW, Wilson K, Pilkington JG, Stear MJ, Pemberton JM

(2001) A microsatellite polymorphism in the gamma interferon

gene is associated with resistance to gastrointestinal nematodes

in a naturally-parasitized population of Soay sheep. Parasitology

122(5):571–582

Cornuet J-M, Luikart G (1996) Description and power analysis of two

tests for detecting recent population bottlenecks from allele

frequency data. Genetics 144(4):2001–2014

Cousins DV, Florisson N (2005) A review of tests available for use in

the diagnosis of tuberculosis in nonbovine species. Rev Sci Tech

Off Int Epiz 24(3):1039–1059

Cross PC, Drewe J, Patrek V, Pearce G, Samuel MD, Delahay RJ

(2009) Host population structure and implications for disease

management. In: Delahay RJ, Smith GC, Hutchings MR (eds)

Management of disease in wild mammals. Springer, Tokyo,

pp 9–30

de Garine-Wichatitsky M, Caron A, Kock R, Tschopp R, Munyeme

M, Hofmeyr M, Michel A (2013) A review of bovine tubercu-

losis at the wildlife-livestock-human interface in sub-Saharan

Africa. Epidemiol Infect 141:1342–1356

Donnelly CA, Woodroffe R, Cox DR, Bourne FJ, Gettinby G,

LeFevre AM, McInerney JP, Morrison WI (2003) Impact of

localized badger culling on TB incidence in British cattle. Nature

426:834–837

Donnelly CA, Woodroffe R, Cox DR, Bourne FJ, Cheeseman CL,

Clifton-Hadley RS, Wei G, Gettinby G, Gilks P, Jenkins H,

Johnston WT, LeFevre AM, McInerney JP, Morrison WI (2006)

Positive and negative effects of widespread culling on tubercu-

losis on cattle. Nature 439(16):843–846

Downing T, Cormican P, O’Farrelly C, Bradley D, Lloyd A (2009)

Evidence of the adaptive evolution of immune genes in chicken.

BMC Res Notes 2(1):254

Ezenwa VO, Etienne RS, Luikart G, Beja-Pereira A, Jolles AE (2010)

Hidden consequences of living in a wormy world: nematode-

induced immune suppression facilitates tuberculosis invasion in

African buffalo. Am Nat 176(5):619–625

Fitzgerald SD, Kaneene JB (2013) Wildlife reservoirs of bovine

tuberculosis worldwide: hosts, pathology, surveillance, and

control. Vet Pathol 50(3):488–499

Frere CH, Prentis PJ, Gilding EK, Mudge AM, Cruickshank A,

Godwin ID (2011) Lack of low frequency variants masks

patterns of non-neutral evolution following domestication. PLoS

One 8:e23041

298 Conserv Genet (2015) 16:289–300

123

http://dx.doi.org/10.1371/journal.pone.0028671
http://dx.doi.org/10.1371/journal.pone.0028671


Goodchild M, Sahu S, Wares F, Dewan P, Shukla RS, Chauhan LS,

Floyd K (2011) A cost-benfit analysis of scaling up tuberculosis

control in India. Int J Tuberc Lung Dis 15(3):358–362

Goudet J (2001) FSTAT: a computer program to calculate F-statistics.

J Hered 86(6):485–486

Groenen MAM, Dijkhof RJM, ban der Peol JJ (1995) Characteriza-

tion of a GlyCAM1-like gene (glycosylation-dependent cell

adhesion molecule 1) which is highly and specifically expressed

in the lactating bovine mammary gland. Gene 158(2):189–195

Harris RB, Wall WA, Allendorf FW (2002) Genetic consequences of

hunting: what do we know and what should we do? Wildl Soc B

30(2):634–643

Hedrick PW (2011) Rapid decrease in horn size of bighorn sheep:

environmental decline, inbreeding depression, or evolutionary

response to trophy hunting? J Hered 2011:1–12

Ihle S, Ravaoarimanana I, Thomas M, Tautz D (2006) An analysis of

signatures of selective sweeps in natural populations of the house

mouse. Mol Biol Evol 23(4):790–797

Jolles AE (2007) Population biology of African buffalo (Syncerus

caffer) at Hluhluwe-iMfolozi Park, South Africa. Afr J Ecol

45:398–406

Jolles AE, Cooper DV, Levin SA (2005) Hidden effects of chronic

tuberculosis in African buffalo. Ecology 86(9):2258–2264

Jolles AE, Etienne RS, Olff H (2006) Independent and competing

disease risks: implications for host populations in variable

environments. Am Nat 167(5):745–757

Jolles AE, Ezenwa VO, Etienne RS, Turner WC, Olff H (unpub-

lished) Interactions between macroparasites and microparasites

drive infection patterns in free-ranging African buffalo

Joseph MB, Mihaljevic JR, Arellano AL, Kueneman JG, Preston DL,

Cross PC, Johnson PTJ (2013) Taming wildlife disease: bridging

the gap between science and management. J Appl Ecol 50(3):

702–712

Kadarmideen HN, Ali AA, Thomson PC, Muller B, Zinsstag J (2011)

Polymorphisms of the SLC11A1 gene and resistance to bovine

tuberculosis in African Zebu cattle. Anim Genet 42(6):656–658

Kappes SM, Keele JW, Stone RT, McGraw RA, Sonstegard TS,

Smith TP, Lopez-Corrales NL, Beattie CW (1997) Genome Res

7(3):235–249

Kimura M, Ohta T (1978) Stepwise mutation model and distribution of

allelic frequencies in a finite population. PNAS 75(6):2868–2872

Koskella B, Lively CM (2009) Evidence for negative frequency-

dependent selection during experimental coevolution of a fresh-

water snail and a sterilizing trematode. Evolution 63(9):2213–

2221

Koskella B, Lin DM, Buckling A, Thompson JN (2012) The costs of

evolving resistance in heterogeneous parasite environments.

Proc R Soc B 279(1735):1896–1903

Koyanagi M, Jerns JA, Chung L, Zhang Y, Moloveanu T, Malik HS,

Bix M (2010) Diversifying selection and functional analysis of

interleukin-4 suggests antagonism-driven evolution at receptor-

binding interfaces. BMC Evol Biol 10:223–236

Lankau RA, Strauss SY (2010) Newly rare or newly common:

evolutionary feedbacks through changes in population density

and relative species abundance, and their management implica-

tions. Evol Appl 4:338–353

Leung TLF, King KC, Wolinska J (2012) Escape from the Red

Queen: an overlooked scenario in coevolutionary studies. Oikos

121(5):641–645

Lewontin RC (1964) The interaction of selection and linkage. I. General

considerations; heterotic models. Genetics 49(1):49–67

Lewontin RC, Kojima K (1960) The evolutionary dynamics of

complex polymorphisms. Evolution 14(4):458–472

Liu K, Muse SV (2005) PowerMarker: an integrated analysis

environment for genetic marker analysis. Bioinformatics 21(9):

2128–2129

Llewellyn D, Thompson MB, Brown GP, Phillips BL, Shine R (2012)

Reduced investment in immune function in invasion-front

populations of the cane toad (Rhinella marina) in Australia.

Biol Invasions 14:999–1008

Luikart G, Cornuet J-M (1998) Empirical evaluation of a test for

detecting recent historical population bottlenecks. Conserv Biol

12:228–237

Maddox JF, Davies KP, Crawford AM, Hulme DJ, Vaiman D, Cribiu

EP, Freking BA, Beh KJ, Cockett NE, Kang N, Riffkin CD,

Drinkwater R, Moore SS, Dodds KG, Lumsden JM, van Stijn

TC, Phua SH, Adelson DL, Burkin HR, Broom JE, Buitkamp J,

Cambridge L, Cushwa WT, Gerard E, Galloway SM, Harrison B,

Hawken RJ, Hiendleder S, Henry HM, Medrano JF, Paterson

KA, Schibler L, Stone RT, van Hest B (2001) An enhanced

linkage map of the sheep genome comprising more than 1000

loci. Genome Res 11(7):1275–1289

Makinen HS, Cano JM, Merila J (2008) Identifying footprints of

directional and balancing selection in marine and freshwater

three-spined stickleback (Gasterosteus aculeatus) populations.

Mol Ecol 17(15):3565–3582

Maruyama T, Fuerst PA (1984) Population bottlenecks and nonequilib-

rium models in population genetics. I. Allele numbers when

populations evolve from zero variability. Genetics 108(3):745–763

McCallum H (2008) Tasmanian devil facial tumour disease: lessons

for conservation biology. TREE 23(11):631–637

McTaggart S, Obbard DJ, Conlon C, Little TJ (2012) Immune genes

undergo more adaptive evolution than non-immune system genes

in Daphnia pulex. BMC Evol Biol 12:63. doi:10.1186/1471-

2148-12-63

Michel AL, Bengis RG, Keet DF, Hofmeyr M, deKlerk LM, Cross

PC, Jolles AE, Cooper D, Whyte IJ, Buss P, Godfroid J (2006)

Wildlife tuberculosis in South African conservation areas:

implications and challenges. Vet Microbiol 112:91–100

Michel AL, Muller B, van Helden PD (2010) Mycobacterium bovis at

the animal-human interface: a problem or not? Vet Microbiol

140:371–381

Myerstud A (2011) Selective harvesting of large mammals: how often

does it result in directional selection? J Appl Ecol 48(4):827–834

Myerstud A, Bischof R (2010) Can compensatory culling offset

undesirable evolutionary consequences of trophy hunting? J Anim

Ecol 79(1):148–160

Nei M (1987) Molecular Evolutionary Genetics. Columbia University

Press, New York

Palermo S, Capra E, Torremorell M, Dolzan M, Davoli R, Haley CS,

Giuffra E (2009) Toll-like receptor 4 genetic diversity among pig

populations. Anim Genet 40(3):289–299

Pollock NR, Campos-Neta A, Kashino S, Napalitano D, Behar SM,

Shin D, Sloutsky A, Joshi S, Guillet J, Wong M, Nardell E

(2008) Discordant quantiFERON-TB Gold tests among US

healthcare workers with increased risk of latent tuberculosis

infection: a problem or solution? Infect Control Hosp Epidemiol

29(9):878–886

Pope LC, Butlin RK, Wilson GJ, Woodroffe R, Erven K, Conyers

CM, Franklin T, Delahay RJ, Cheeseman CL, Burke T (2007)

Genetic evidence that culling increases badger movement:

implications for the spread of bovine tuberculosis. Mol Ecol

16(23):4919–4929

Raymond M, Rousset F (1995) GENEPOP: population genetics

software for exact tests and ecumenicism. J Hered 86(3):248–249

Renehan AG, Zwahlen M, Minder C, O’Dwyer ST, Shalet SM, Egger

M (2004) Insulin-like growth (IGF)-I, IGF binding protein-3,

and cancer risk: systematic review and meta-regression analysis.

Lancet 363(19):1346–1353

Renwick AR, White PCL, Bengis RG (2007) Bovine tuberculosis in

southern African wildlife: multi-species host-pathogen system.

Epidemiol Infect 135(4):529–540

Conserv Genet (2015) 16:289–300 299

123

http://dx.doi.org/10.1186/1471-2148-12-63
http://dx.doi.org/10.1186/1471-2148-12-63


Robinson SJ, Samuel MD, O’Rourke KI, Johnson CJ (2012) The role

of genetics in chronic wasting disease of North American

cervids. Prion 6(2):153–162

Ryan TJ, Buddle BM, de Lisle GW (2000) An evaluation of the

gamma interferon test for detecting tuberculosis in cattle 8 to

28 days after tuberculin skin testing. Res Vet Sci 69:57–61

Sackett LC, Collinge SK, Martin AP (2013) Do pathogens reduce

genetic diversity of their hosts? Variable effects of sylvatic

plague in black-tailed prairie dogs. Mol Ecol 22(9):2441–2455

Schneider S, Roessli D, Excoffier L (2000) Arlequin: an integrated

software package for population genetics data analysis. Evol

Bioinform 1:47

Shim E, Galvani AP (2009) Evolutionary repercussions of avian

culling on host resistance and influenza virulence. PLoS One

4(5):e5503

Singleton GR, Brown PR, Jacob J, Aplin KP, Sudaramaji (2007)

Unwanted and unintended effects of culling: a case for

ecologically-based rodent management. Integr Zool 2:247–257

Smith KF, Acevedo-Whitehouse K, Pederson AB (2009) The role of

infectious diseases in biological conservation. Anim Conserv

12:1–12

Sommer S (2005) The importance of immune gene variability (MHC)

in evolutionary ecology and conservation. Front Zool 2:16–34

Spurgin LG, Richardson DS (2010) How pathogens drive genetic

diversity: MHC, mechanisms and misunderstandings. Proc R

Soc B 277:979–988

Thrall PJ, Burdon JJ (2003) Evolution of virulence in a plant host-

pathogen metapopulation. Science 299:1735–1737

Tocyap ML, Azar N, Chen T, Wiggs J (2006) Clinical and molecular

characterization of a patient with an interstitial deletion of

chromosome 12q15-q23 and peripheral corneal abnormalities.

Am J Opthalmol 141(3):566–567

Tonteri A, Vasemagi A, Lumme J, Primmer C (2010) Beyond MHC:

signals of elevated selection pressure on Atlantic salmon (Salmo

salar) immune-relevant loci. Mol Ecol 19(7):1273–1282

Tschirren B, Raberg L, Westerdahl H (2011) Signatures of selection

acting on the innate immunity gene Toll-like receptor 2 (TLR2)

during the evolutionary history of rodents. J Evol Biol 24(6):

1232–1240

van Oosterhout C, Hutchinson W, Wills D, Shipley P (2004) MICRO-

CHECKER: software for identifying and correcting genotyping

errors in microsatellite data. Mol Ecol Notes 4(3):535–538

Waters WR, Tacker TC, Nonnecke BJ, Palmer MV, Schiller I, Oesch

B, Vordermeier HM, Silva E, Estes DM (2012) Evaluation of

Gamma Interferon (IFN-c)-Induced protein 10 responses for
detection of cattle infected with Mycobaterium bovis: compar-

isons to IFN-c responses. Clin Vaccine Immunol 19(3):346–351
White PJ, Wallen RL, Geremia C, Treanor JJ, Blanton DW (2011)

Management of Yellowstone bison and brucellosis transmission

risk—implications for conservation and restoration. Biol Con-

serv 144(5):1322–1334

Wood PR, Corner LA, Rotherl JS, Baldock C, Jones SL, Cousins DB,

McCormick BS, Francis BR, Creeper J, Tweddle NE (1991)

Field comparison of the interferon-gamma assay and the

intradermal tuberculin test for the diagnosis of bovine tubercu-

losis. Aust Vet J 68(9):286–290

Woodroffe R (1999) Managing disease threats to wild mammals.

Anim Conserv 2:185–193

Woodroffe R, Donnelly CA, Cox DR, Bourne FJ, Cheeseman CL,

Delahay RJ, Gettinby G, McInerney JP, Morrison WI (2006)

Effects of culling on badger Meles meles spatial organization:

implications for the control of bovine tuberculosis. J Appl Ecol

43(1):1–10

Woodroffe R, Donnelly CA, Wei G, Cox DR, Bourne FJ, Burke T,

Butlin RK, Cheeseman CL, Gettinby G, Gilks P, Heges S,

Jenkins HE, Johnston WT, McInerney JP, Morrison WI, Pope

LC (2009a) Social group size affects Mycobacterium bovis

infection in European badgers (Meles meles). J Anim Ecol

78:818–827

Woodroffe R, Donnelly CA, Cox DR, Gilks P, Jenkins HE, Johnston

WT, LeFevre AM, Bourne FJ, Cheeseman CL, Clifton-Hadley RS,

Gettinby G, Hewinson RG, McInerney JP, Mitchell AP, Morrison

WI, Watkins GH (2009b) Bovine tuberculosis in cattle and badgers

in localized culling areas. J Wildl Dis 45(1):128–143

Yang F, Eckardt S, Leu NA, McLaughlin KJ, Wang PJ (2008) Mouse

TEX15 is essential for double-strand break repair and chromo-

somal synapse during male meiosis. J Cell Biol 180(4):673–679

300 Conserv Genet (2015) 16:289–300

123


	Signatures of natural and unnatural selection: evidence from an immune system gene in African buffalo
	Abstract
	Introduction
	Methods
	Sample collection
	Molecular methods
	Locus descriptions
	Indices of genetic diversity and GD
	Patterns of selection
	Genetic diversity in HIP

	Results
	Locus descriptions
	Signature of selection: patterns of genetic diversity and GD
	Reduced genetic diversity in HIP

	Discussion
	Evidence of selection in KNP
	Management-driven selection in HIP

	Conclusions
	Acknowledgments
	References