This repo contains the data in relation to the maxATAC project. For complete installation directions of maxATAC please refer to the maxATAC repo, https://github.com/MiraldiLab/maxATAC
This directory contains:
hg38_maxatac_blacklist.bed: bed file of maxATAC blacklist regionshg38_maxatac_blacklist.bw: bed file of maxATAC blacklist regionshg38.chrom.sizes: hg38 chromosome sizes file
Curation of the maxATAC black list is described in the maxATAC preprint
Information for the 127 TF models available through maxATAC. Each directory is a TF name. Each directory contains a .h5 of the best TF model, a .tsv file of threshold statistics, and a .png of the statistics.
Example directory structure:
|____ATF2
| |____ATF2_validationPerformance_vs_thresholdCalibration.png
| |____ATF2_binary_revcomp99_fullModel_RR0_73.h5
| |____ATF2_validationPerformance_vs_thresholdCalibration.tsv
By default, the threshold for calling TFBS is set to maximize the validation F1-score of the TF model. The threshold statistics file and image enables users to see the trade-offs between precision and recall for each model, information that could guide their parameter choices for maxatac predict
This is an example of the first 6 lines of the threshold statistics file used to call peaks. The first column is the threshold. Every column after is the average performance reported for different metrics on validation chromosome 2.
Standard_Thresh Monotonic_Avg_Precision Monotonic_Avg_Recall Monotonic_Avg_log2FC Avg_F1
0.0 0.0035793736711456103 0.9910269020975186 0.4617435633366975 0.0071224011910284366
0.001 0.018103716393149 0.9242690259191988 2.996816276717381 0.03517534734171995
0.002 0.027523721672056148 0.92192080328679 3.5979115724538913 0.05275399945034204
0.003 0.03384312047545328 0.9123870842533509 3.8788363137406194 0.06429293366464785
0.004 0.03872007478681639 0.9030209327592842 4.055800482157333 0.07305708818396862
0.005 0.04270041191717275 0.9017292649083462 4.199724330597559 0.0801251221410759
A directory of scripts that are used to process the ATAC-seq, DHS, and ChIP-seq data analyzed by our publication. The scripts ATAC_bowtie2_pipeline.sh and scatac_generate_bigwig.sh are required for running maxatac prepare.
.
|____ATAC
| |____align_reads_STAR.sh
| |____infer_insertion_sites.sh
| |____ATAC_bowtie2_nodedup_pipeline.sh
| |____RPM_normalize_Tn5_counts.sh
| |____shift_reads.sh
| |____trim_reads.sh
| |____infer_Tn5_sites.sh
| |____scatac_generate_bigwig.sh
| |____ATAC_bowtie2_pipeline.sh
| |____sra2fastq_PE.sh
| |____filter_BAM_STAR.sh
| |____ATAC_post_alignment.sf
|____DHS
| |____infer_DHS_cut_sites.sh
| |____generate_DHS_coverage_bigwig.sh
| |____filter_DHS_SE_ENCODE.sh
|____CHIP
| |____infer_5prime_BAM.sh
| |____filter_chroms_from_bam.sh
| |____slop_5prime.sh
| |____RPM_normalize_5prime_counts.sh
This .tsv file shows an example meta file required to train your own models.
The directory of motif information from CISBP. There are two meta files, (1) a list of all TF motifs that have been curated and (2) metadata about family source, etc... from CISBP.
.
|____20201001_all_motifs_meta.tsv
|____CISBP_2_TF_Information_all_motifs_plus.txt
|____all_motifs
| |____M00008_2.00.pfm
| |____M00009_2.00.pfm
| |____...
The maxATAC pre-print is currently available on bioRxiv.
maxATAC: genome-scale transcription-factor binding prediction from ATAC-seq with deep neural networks
Tareian Cazares, Faiz W. Rizvi, Balaji Iyer, Xiaoting Chen, Michael Kotliar, Joseph A. Wayman, Anthony Bejjani, Omer Donmez, Benjamin Wronowski, Sreeja Parameswaran, Leah C. Kottyan, Artem Barski, Matthew T. Weirauch, VB Surya Prasath, Emily R. Miraldi
bioRxiv 2022.01.28.478235; doi: https://doi.org/10.1101/2022.01.28.478235
