key: cord-0761160-rdh2ni19
authors: Halfmann, P.; Minor, N. R.; Haddock, L. A.; Maddox, R.; Moreno, G. K.; Braun, K.; Baker, D.; Riemersma, K.; Prasad, A.; Alman, K. J.; Lambert, M. C.; Florek, K.; Bateman, A.; Westergaard, R.; Safdar, N.; Andes, D. R.; Kawaoka, Y.; Fida, M.; Yao, J. D.; Friedrich, T.; O'Connor, D. H.
title: Evolution of a globally unique SARS-CoV-2 Spike E484T monoclonal antibody escape mutation in a persistently infected, immunocompromised individual.
date: 2022-04-12
journal: nan
DOI: 10.1101/2022.04.11.22272784
sha: ccb7adaee55e8bac1874575c7b547baad685570c
doc_id: 761160
cord_uid: rdh2ni19

Prolonged infections in immunocompromised individuals may be a source for novel SARS-CoV-2 variants, particularly when both the immune system and antiviral therapy fail to clear the infection, thereby promoting adaptation. Here we describe an approximately 16-month case of SARS-CoV-2 infection in an immunocompromised individual. Following monotherapy with the monoclonal antibody Bamlanivimab, the individual's virus was resistant to this antibody via a globally unique Spike amino acid variant (E484T) that evolved from E484A earlier in infection. With the emergence and spread of the Omicron Variant of Concern, which also contains Spike E484A, E484T may arise again as an antibody-resistant derivative of E484A.

The host immune response typically clears SARS-CoV-2 infections by 15 to 25 days [1] [2] [3] [4] , and this short duration of active virus replication limits within-host evolution. In contrast, infections in immunocompromised individuals can be prolonged, with the most prolonged previously documented infection lasting 333 days 5 . Prolonged infections may facilitate SARS-CoV-2 evolution by providing time for within-host selection to act, potentially driving the emergence of immunologic escape variants (reviewed in 6 ).

Spike residue E484 lies in the receptor-binding domain (RBD), which is the target for approximately 90% of neutralizing antibodies 7 . Class 2 Spike neutralizing antibodies, as defined in Barnes et al. 2020 , are particularly sensitive to Spike residue 484 substitutions 8 . Class 2 neutralizing antibodies form a hydrogen bond network with Spike residue 484, which is highly conserved across class 2 antibodies. A recent study exposed a VSV-eGFP-SARS-CoV-2-S chimeric virus to 19 neutralizing monoclonal antibodies (mAbs) and found that substitutions appeared at E484 more frequently than at any other residue. E484A, E484D, E484G, and E484K all reduce the neutralization activity of antibodies in convalescent plasma 9 . Spike E484K is the best studied of these variants; the substitution was found to reduce susceptibility to the monoclonal antibody Bamlanivimab by more than 2,360-fold 10 . 11 . Providing further support for this hypothesis, E484A variants have also arisen in immunocompromised individuals with prolonged infection 11, 12 . SARS-CoV-2 culture in the presence of monoclonal antibodies can also select E484A in vitro 13 . E484A was transiently detected in one patient receiving Bamlanivimab monotherapy, but the impact of this mutation on Bamlanivimab recognition was not directly tested 14 .

Here, we report an ongoing, 16-month-long SARS-CoV-2 infection in an immunocompromised patient.

Following unsuccessful Bamlanivimab monotherapy, a number of substitutions were detected in the consensus sequence of the infecting virus, including one encoding Spike E484T, a substitution that has not been reported previously in public SARS-CoV-2 sequence databases. Virus isolated from this patient encoding Spike E484T was highly resistant to Bamlanivimab, but remained susceptible to three other commercial monoclonal antibodies. These findings highlight the necessity of frequent testing, viral sequencing, and proactive treatment for prolonged SARS-CoV-2 infections occurring in immunocompromised individuals.

The patient described in this case is in their 50s and is immunocompromised. In Fall 2020, the patient reported sinus congestion and pain breathing to their primary care provider, and a nasopharyngeal swab specimen first tested PCR-positive for SARS-CoV-2. The patient's symptoms worsened into fever, shortness of breath, and body fatigue by ~two months after their initial diagnosis. The patient was admitted to a hospital emergency department, where physicians found evidence for organizing pneumonia and alveolar damage, and no evidence of bacterial or fungal infection. Despite hypoxia and recurrent fevers, the patient's blood oxygen saturation remained in the healthy range above 90 mmHg, and they were sent home. There, the patient started a course of dexamethasone. Their fevers improved, but their hypoxia worsened, requiring supplemental oxygen via nasal cannula during sleep and activity. The patient's blood oxygen saturation continued to fall into the 70s, requiring placement on a bilevel positive airway pressure (BiPAP) ventilator to bring oxygenation back into the 90s.

About four months post-diagnosis, the patient was admitted to a hospital Intensive Care Unit (ICU) and immediately intubated. The patient was extubated less than a week later, though they were unable to wean off supplemental oxygen via nasal cannula. A nasopharyngeal swab specimen from the patient once again tested PCR-positive for SARS-CoV-2, which prompted the ICU physicians to administer two courses of inpatient remdesivir, two units of convalescent plasma, and an aggressive steroid regimen. The physicians also requested compassionate use of monoclonal antibodies from Eli Lilly and Regeneron, though at the time, both companies denied their requests. Yet another nasopharyngeal swab PCR test came back positive for SARS-CoV-2. The patient received intravenous immunoglobulin, and less than a week later the patient was discharged on a steroid regimen and supplemental oxygen via nasal cannula.

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.11.22272784 doi: medRxiv preprint Around six months post-diagnosis, the physicians submitted another request for compassionate use of a monoclonal antibody therapy to Eli Lilly. This time, the company approved their request, and the patient received 700mg of intravenous Bamlanivumab. More than a month later, the patient reported worsening shortness of breath and low-grade fever, and chest CT scans again showed organizing pneumonia. Additional nasopharyngeal swabs came back positive, though without any major changes in symptoms and pulmonary function. Eventually, at almost 13 months post-diagnosis, the patient's cough and difficulty breathing worsened again, at which point they sought a second opinion from another facility. After another PCR-positive nasopharyngeal swab, the physicians administered three Table 1 ).

At the time of this writing, it remains unclear whether the patient is experiencing symptoms from a persistent, actively replicating viral infection, or is suffering from post-COVID lung damage with persistent viral excretion in the respiratory tract.

Note that the above case history has been abbreviated, with certain delails withheld at the request of medRxiv. The full case description and timeline is available upon reasonable request to the corresponding author.

Prolonged detection of SARS-CoV-2 by RT-qPCR and virus culture Ct values were below 27 in the majority of nasopharyngeal swab PCR tests, indicating active viral replication and high viral loads ( Figure 1A ). At nearly ten months post-diagnosis, infectious viruses from . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 12, 2022. ;  a nasopharyngeal swab were successfully cultured in vitro. Additionally, viral isolates from the immunocompromised individual were sequenced at twelve timepoints, beginning at ~4 months post-diagnosis. We found that consensus sequences from each of these timepoints are more closely related to one another than any other sequences in GISAID and are all Pango lineage B.1.2, consistent with a prolonged infection rather than reinfection ( Figure 2B ). To our knowledge, this presumptive 16-month ongoing infection is the most prolonged SARS-CoV-2 infection reported to date, and it highlights the need for an arsenal of medical interventions in persistently infected immunocompromised individuals. Additionally, the virus described in this case accumulated enough mutations within its host to roughly match the global SARS-CoV-2 mutation rate for the whole of the pandemic (see Supplemental Figure   2 ).

A single, prolonged infection is the most parsimonious explanation for these findings. However, a household member tested positive for SARS-CoV-2 about three months post-diagnosis, and it is theoretically possible that the immunocompromised individual cleared their infection between the previously confirmed PCR test on day 0 and was reinfected by their household contact. Unfortunately, no samples from this household member are available to investigate this possibility.

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.11.22272784 doi: medRxiv preprint A.

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.11.22272784 doi: medRxiv preprint Emergence of the globally unique Spike E484T and other mutations in response to Bamlanivumab At the earliest timepoint for which sequencing data is available, taken about four months into the infection, we detected 24 mutations that separated the patient's virus from Wuhan-1. This number gradually increased to 40 mutations by our final sequencing timepoint. Early mutations included Envelope T30I, one of the most frequently arising mutations immunocompromised individuals 15 , though this mutation was lost by approximately five months post-diagnosis. Also in our earliest sequencing timepoint, we detected a point mutation in Spike codon 484 (GAA>GCA), which changed the amino acid glutamic acid (E) to alanine (A). This Spike E484A mutation persisted in the virus consensus sequence for five further timepoints, at least to eight months post-diagnosis. At that point, the patient received the Bamlanivumab monoclonal antibody therapy, and their virus was subsequently not sequenced for more than three months; there was no additional testing or sequencing during this interval. At the next point at which their virus was sequenced, we observed a cumulative total of 37 consensus mutations from the Wuhan-1 sequence, eleven of which had risen together from sub-consensus variants detected at previous timepoints. One of the eleven that arose together was an additional, novel point mutation in the Spike codon 484 (GCA>ACA), which changed the amino acid residue to threonine (T).

As of February 2022, Spike E484T has not been documented in any other SARS-CoV-2 samples. The substitution remained in our next sequencing timepoint at 11 months post-diagnosis. However, by the following timepoint approximately two months later, the nucleotide mutation that caused E484T had returned to a frequency far below consensus, as did ten mutations that rose with it. The virus consensus sequence at ~13 months post-diagnosis, Spike codon 484 had reverted to E484A (Supplemental . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.11.22272784 doi: medRxiv preprint

The degree of genetic evolution we document in this patient's virus highlights the potential role of immunocompromised individuals in generating and propagating novel variants. We also describe the first case of a globally unique Spike E484T variant, which arose after treatment with the Bamlanivumab monotherapy. While E484T appears to be novel for SARS-CoV-2, the bat betacoronavirus RaTG13, one of the most similar coronaviruses to SARS-CoV-2 in terms of nucleotide identity, also bears a threonine at Spike codon 484. This suggests that E484T is a variant that can be tolerated in betacoronaviruses with similar Spike structure to SARS-CoV-2, though the RaTG13 Spike has noteworthy differences, including lower stability in its pre-furin-cleaved form 16 . Additionally, in silico E484T mutants of the SARS-CoV-2 Spike protein showed increased transduction of mouse ACE2 receptors, the key for viral entry into the cell, by more than 16-fold 17 .

In this case, it is unfortunate-though certainly not unprecedented-that residual specimens from this patient's earliest SARS-CoV-2 tests are not available for retrospective analyses. Given the volume of specimen required for RT-qPCR, most providers discard residual nasopharyngeal swab specimens after testing. However, this case highlights the potential value of archiving test-positive specimens from Although data are limited in the county where this individual lives, surveillance sequencing suggests that there was no onward spread of the individual's virus, notwithstanding the household case of unknown provenance. This lack of spread is particularly consequential for the Bamlanivumab-evading virus, which bore the novel Spike E484T along with a cumulative 36 other mutations. That E484T is unique to this individual raises the possibility that E484T, and the other mutations that emerged and disappeared along with it (Supplemental Figure 1) , has an associated fitness cost.

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.11.22272784 doi: medRxiv preprint Another, non-mutually exclusive explanation for the lack of onward spread is that the patient's collaboration with state and local public health officials has been successful at managing their infection, especially after Spike E484T was detected, and indeed it is to their credit that this case report involves only one individual. That individual has remained isolated at home for the majority of their 16-month, ongoing infection, and has cooperated with ongoing monitoring efforts by state public health officials. What support is available to these individuals in the event that they cannot maintain employment or afford healthcare? How do therapeutics intended to manage symptoms-or the lack thereof-influence the virus's evolution? These questions echo previous controversies, such as the isolation of individuals with highly drug-resistant tuberculosis 22 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.11.22272784 doi: medRxiv preprint

Collection and testing of biological specimens and protected health information were performed in concordance with the University of Wisconsin IRB # 2021-0076 following informed consent from the patient. Table 1 ).

To obtain a semi-quantitative measurement of the ~10-month virus, we isolated viral RNA from an additional nasopharyngeal swab sample using the Total Nucleic Acid kit for the Maxwell RSC instrument (Promega, Madison, WI). We used the N1 qRT-PCR assay developed by the CDC and commer- is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.11.22272784 doi: medRxiv preprint (PFU). Plates were incubated at 37ºC for 30 minutes. Culture supernatant was aspirated from Vero E6/TMPRSS2 cells plated in tissue culture 96-plates and replaced with the mixtures of serially diluted antibody and virus mixture (100µl/well) followed by incubation at 37°C with 5% CO2. The inhibitor concentration that prevented 99% cell death (IC99 value) at the given concentration (as determined by cytopathic effects visualized by a brightfield microscope) defines antibody potency.

We performed sequencing on a nasopharyngeal or nasal swab sample that was leftover after twelve of the immunocompromised individual's PCR tests. All viral RNA extractions over the course of this study took place on the Maxwell RSC Viral Total Nucleic Acid Purification Kit (Promega, USA), used according to the manufacturer's instructions.

To generate sequence data, we first sequenced viral cDNA libraries on an Oxford Nanopore (ONT) min-ION with the ARTIC v3 amplicon approach (https://www.protocols.io/view/covid-19-artic-v3-illumina-library-construction-an-bibtkann), but the previously documented amplicon dropout near the beginning of the Spike gene occurred in all of our samples. To improve breadth of coverage across the whole of the Spike gene, we re-sequenced ten of the samples on an ONT MinION using the MIDNIGHT protocol (https://www.protocols.io/view/sars-cov2-genome-sequencing-protocol-1200bp-amplic-bwyppfvn) 25 .

To examine minor variant frequencies with maximum breadth of coverage, we later used the ARTIC v4.1 protocol to sequence an overlapping ten samples on an Illumina MiSeq (https://www.protocols.io/ view/sars-cov-2-sequencing-on-illumina-miseq-using-arti-bfefjjbn).

Preparation for ARTIC v4.1 sequencing began with reverse transcription, for which we combined 11 µl of leftover viral isolate from ten of the samples with 4µl of SSIV Vilo Master Mix (Invitrogen, USA) and 5µl of DEPC water. We then incubated these pools through 25°C for 10 minutes, 50°C for 10 minutes, then 85°C for 5 minutes, with a final indefinite hold of 4°C. Next, we amplified 2.5µl of cDNA in two independent multiplexed PCR reactions with 12.5µl of Phusion High Fidelity Master Mix (New England BioLabs), 6µl of DEPC water and 4µl of ARTIC primer pools. After an initial 30 second denaturation at 98°C, we amplified the cDNA replicates over 30 cycles of 98°C for 30 seconds, 98°C for 15 seconds, and 65°C for 5 min, with an indefinite hold at 4°C to finish. We then combined and purified the amplicon pools with a 1:1 concentration of AMPure XP beads (Beckman Coulter, USA), and then quantified them using the Qubit hsDNA kit (Invitrogen, USA). For those samples being sequenced on the Illumina MiSeq (see Supplemental Table 1 ), we diluted all samples to 150ng in 60µl for use as inputs for the . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.11.22272784 doi: medRxiv preprint TruSeq DNA Nano Kit (Illumina, USA). The final, purified amplicon pools were then sequenced with the MiSeq reagent v3-600 Kit at 10pM with 1% PhiX. For those samples being sequenced with ARTIC on an ONT minION (see Supplemental Table 1 ), we mixed 12 µl of the purified DNA library, 37.5 µl of the ONT Sequencing Buffer II, and 25.5 µl of ONT Loading Solution, and loaded the mixture onto the MinION for sequencing. Additionally, our only sequence data from our timepoint at approximately 11 months post-diagnosis was sequenced on an Illumina MiSeq using an ARTIC v3 amplicon approach that was very similar to that described above, except that it uses tagmentation to fragment and barcode viral vDNA prior to sequencing (https://www.protocols.io/view/sars-cov-2-sequencing-on-illumina-miseq-using-arti-n92ld9w1xg5b/v1) 26 .

To reverse transcribe the viral isolate in preparation for sequencing with the MIDNIGHT protocol, we combined 2 µl of LunaScript RT SuperMix (New England BioLabs, USA) with 8 µl of sample, and then incubated the pools at 25°C for 2 minutes, 55°C for 10 minutes, 95°C for 1 minute, and a final hold at 4°C. Next, we combined 2.5 μl of each incubated isolate together into one of two pools (pool A and pool B), both with 3.7 µl of nuclease-free water, 0.05 µl of MIDNIGHT primer pool, and 6.25 µl of Q5 HS Master Mix (New England BioLabs). We then amplified this mix with an initial 30 second denaturation at 98°C followed by 35 cycles of 98°C for 15 seconds, 61°C for 2 mins, and 65°C for 3 mins, with an indefinite 4°C hold to finish. 5 µl from each of these PCR product pools were then mixed with 2.5 μl of sample-specific barcode and 2.5ul of nuclease-free water, multiplexed, purified with 1X SPRI beads (Beckman Coulter, USA), and quantified with the Qubit dsDNA HS Assay Kit. Finally, a mixture of 12 µl of the purified DNA library, containing approximate 800 ng of DNA, along with 37.5 µl of the ONT Sequencing Buffer II, and 25.5 µl of ONT Loading Solution were loaded onto the ONT MinION Instrument for sequencing. The full MIDNIGHT protocol is available at https://www.protocols.io/view/ sars-cov2-genome-sequencing-protocol-1200bp-amplic-bwyppfvn 25 .

Processing raw sequence data for consensus sequence analyses The majority of analyses described here use consensus sequences, which we assembled for each sample using a workflow customized to run remotely with compute resources from the UW-Madison Center For High Throughput Computing (https://chtc.cs.wisc.edu/), and using the workload manager HTC Condor. Briefly, our SARS-CoV-2 sequence analysis pipeline base-calls reads with the Guppy basecaller. It then trims adapters and primers, maps to the SARS-CoV-2 Wuhan-1 sequence . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.11.22272784 doi: medRxiv preprint (GenBank: MN908947.3), and removes reads that map to more than one location. From there, consensus sequences are assembled with the requirement that each base is present in at least 50% of the downsampled reads. Any bases with a depth of less than 20 and/or an allele fraction under 50% reads are assigned an N. The pipeline then formats consensus sequences and metadata for GISAID upload, where we have deposited all twelve consensus sequences (see Supplemental Table 1 ). In the separate workflow produced specifically for this study, we also call variants from MN908947.3 in these consensus sequences using iVar, which produces simpler outputs than the standard VCF, making them more straightforward to plot in R 27 .

To visualize amplicon dropouts in these consensus sequences, we returned to the ONT MIDNIGHT reads in FASTQ format, which were originally used to generate consensus sequences. First, we mapped the reads to MN908947.3 using minimap2's 'map-ont' preset, clipped reads down to MIDNIGHT amplicon pileups using SAMtools, and then used covtobed to produce BED-format tables of regions where read depth-of-coverage was below 20 [28] [29] [30] . Finally, we used an R script to plot mutations in the consensus sequences through time, along with dropouts.

While our sequence analysis pipeline is most refined for MIDNIGHT ONT data, it comes with the necessary caveat that ONT sequencing instruments have a relatively high error rate. This error rate is unproblematic when generating consensus sequences, but can also obscure low-frequency variants when examining intrahost variant frequencies. To overcome this obstacle, we sequenced leftover samples from ten timepoints on an Illumina MiSeq as described above. We then mapped these reads to MN908947.3 using minimap2's 'sr' preset 29 , clipped reads down to ARTIC amplicon pileups using SAMtools 28 , and called variants in standard VCF format using the BBTools `callvariants.sh` script 31 .

We then annotated protein effects onto these VCFs using snpEff 32 . From there, variants that rose into consensus frequency after Bamlanivumab treatment were plotted using an R script.

To examine the rate at which the patient's virus accumulated mutations, and compare that rate with SARS-CoV-2 globally, we randomly subsampled approximately 5,000 SARS-CoV-2 sequences 33 from the GISAID global dataset using scripts made available in the Nextstrain command line interface 34 . We then mapped the GISAID consensus sequences to MN908947.3 using minimap2, called variants from . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.11.22272784 doi: medRxiv preprint all sequence alignments using the BBTools `callvariants.sh` script, and then used R to compute genetic distance from MN908947.3 and plot.

All scripts, consensus sequence data, and sequencing read data used in the writing of this manuscript can be found at https://github.com/dholab/E484T-visualizations. Additional data from intermediate stages of our analysis are available upon reasonable request to the authors. 

Understanding the outcomes of COVID-19 -does the current model of an acute respiratory infection really fit?

Comparisons of viral shedding time of SARS-CoV-2 of different samples in ICU and non-ICU patients

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in

China: a retrospective cohort study

Asymptomatic infection and atypical manifestations of COVID-19: Comparison of viral shedding duration

Persistent Severe Acute Respiratory Syndrome Coronavirus 2 Infection in

Immunocompromised Host Displaying Treatment Induced Viral Evolution

SARS-CoV-2 variants, spike mutations and immune escape

Mapping neutralizing and immunodominant sites on the SARS-CoV-2 spike receptor-binding domain by structure-guided high-resolution serology

SARS-CoV-2 neutralizing antibody structures inform therapeutic strategies

Identification of SARS-CoV-2 spike mutations that attenuate monoclonal and serum antibody neutralization

Emergency Use Authorization (EUA) for bamlanivimab 700mg IV Center for Drug Evaluation and Research (CDER) Memorandum on Fact Sheet Update

SARS-CoV-2 evolved during advanced HIV disease immunosuppression has

Beta-like escape of vaccine and Delta infection elicited immunity

Persistence and Evolution of SARS-CoV-2 in an Immunocompromised Host

Molecular determinants and mechanism for antibody cocktail preventing SARS-14

Emergence of E484K Mutation Following Bamlanivimab Monotherapy among High-Risk Patients Infected with the Alpha Variant of SARS-CoV-2

Recurrent SARS-CoV-2 Mutations in Immunodeficient Patients

SARS-CoV-2 and bat RaTG13 spike glycoprotein structures inform on virus evolution and furin-cleavage effects

The Rhinolophus affinis bat ACE2 and multiple animal orthologs are functional receptors for bat coronavirus RaTG13 and SARS-CoV-2

SARS-CoV-2 evolution during treatment of chronic infection

Case Study: Prolonged Infectious SARS-CoV-2 Shedding from an

Within-host evolution of SARS-CoV-2 in an immunosuppressed COVID-19 patient: a source of immune escape variants

Prevalence of Immunosuppressive Drug Use Among Commercially Insured US Adults

Legal power and legal rights--isolation and quarantine in the case of drug-resistant tuberculosis

Syrian hamsters as a small animal model for SARS-CoV-2 infection and countermeasure development

Characterization of a new SARS-CoV-2 variant that emerged in Brazil

Rapid and inexpensive whole-genome sequencing of SARS-CoV-2 using 1200 bp tiled amplicons and Oxford Nanopore Rapid Barcoding

A rapid, cost-effective tailed amplicon method for sequencing SARS-CoV-2

An amplicon-based sequencing framework for accurately measuring intrahost virus diversity using PrimalSeq and iVar

Twelve years of SAMtools and BCFtools

Minimap2: pairwise alignment for nucleotide sequences

covtobed: a simple and fast tool to extract coverage tracks from BAM files

BBMap

A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3

GISAID's Role in Pandemic Response

Nextstrain: real-time tracking of pathogen evolution

International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity