TAMC (Transcriptional factor binding prediction from ATAC-seq profile at Motif-predicted binding sites using Conventional neural networks) is an open source tool for predicting motif-centric TF binding activity from paired-end ATAC-seq data. TAMC takes advantage of signal processing strategies in HINT-ATAC and TOBIAS to prepare input signals and make predictions of binding probability using a 1D-conventional neural network (1D-CNN) model.
Python (3.6); Numpy; pyBigWig; pysam; pandas; PyTorch; scipy.stats; math; reg-gen;
git clone https://github.com/tianqiyy/TAMC.git
$ python inputsignal.py --input_format default \
--refgenome reference_genome_directory \
--atac_bam atac-seq.bam \
--TOBIAS_FPS_bw TOBIAS_footprint_score.bigwig \
--mpbs_bed mpbs.bed \
--biastable_F none_bias_table_F.txt \
--biastable_R none_bias_table_R.txt \
--outdir output_directory \
--prefix file_name_prefix_string
TOBIAS_FPS_bw: TOBIAS_footprint_score.bigwig is generated using ATACorrect and FootprintScores tools in TOBIAS package.
$ python train.py --modelname default \
--traindatadir training_input_signal_directory \
--valdatadir validating_input_signal_directory \
--batchsize 35 \
--epochnumber 10 \
--learnrate 0.0001 \
--trainrecorddir training_record_directory \
--bestmodeldir bestmodel_record_directory \
--prefix file_name_prefix_string
We have provided sample trained models for CTCF and EGR1 under the bestmodels folder.
$ python predict.py --modelname default \
--refgenome reference_genome_directory \
--atac_bam atac-seq.bam \
--TOBIAS_FPS_bw TOBIAS_footprint_score.bigwig \
--mpbs_bed mpbs.bed \
--biastable_F none_bias_table_F.txt \
--biastable_R none_bias_table_R.txt \
--bestmodel1 best_trained_model_1.pt \
--bestmodel2 best_trained_model_2.pt \
--bestmodel3 best_trained_model_3.pt \
--outdir output_directory \
--prefix file_name_prefix_string