id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
10_1101-2021_02_10_430606	Wei, Zheng	NeuronMotif: Deciphering transcriptional cis-regulatory codes from deep neural networks	2021	31	.pdf	application/pdf	12013	1107	63	Each point is a decoupled motif generate by a sample set of sequence. Only the max activation value of the decoupled motifs in Fig. 3b are significantly higher than the decoupled motifs of other neurons in layer 3 of Basset-3 model. discovered (q-value < 0.001) from the neuron in convolutional output layer of Basset, BD-5 and BD-10 model. c, The number of motif discovered (q-value < 0.01) from the neuron in layer 3 of Basset model using different sub-patterns in the input feature map of the max pooling layer to split the sequences set of which are DNA-sequence based DCNN models with 3 general convolutional layers for stacking sequences of different synonymous motifs with the maximum activation value In summary, we presented NeuronMotif as an effective algorithm to reveal the cisregulatory motif grammar learned by DCNN model that use DNA sequence to annotate sequences indicate more synonymous motif mixture in this DCNN model.	./cache/10_1101-2021_02_10_430606.pdf	./txt/10_1101-2021_02_10_430606.txt