RoboATAC

This repository accompanies the preprint Liu et al, bioRxiv 2025b.

Code

• code
  • utils: helper functions for motif analysis, ChromBPNet motif clustering, and color palette
  • 00-techdev
    • 01: comparing OmniATAC and RoboATAC QC metrics from benchmarking datasets
    • 02: script to check fraction of reads in peaks
  • 01-preprocessing
    • Using the SE branch of the snakeATAC preprocessing pipeline
    • Snakefile.py: snakemake file to process single-ended Ultima reads of RoboATAC libraries
  • 02-atac
    • 01: create ChrAccR object, consensus peak calling, read normalization, chromVAR
    • 02: dimensionality reduction with PCA
    • 03: matching sequences in consensus peaks to JASPAR2020 motifs
    • 04: differential peak analysis
    • 05: correlating ATAC and RNA differentials
    • 06: TF footprinting
    • 07: dimensionality reduction with UMAP
    • 08: analyze peak type compositions of differential peak sets
    • 09: motif scores and motif counts within differential peak sets
    • 10: linear and Hill fits of peak dose response to determine peak sensitivity group
    • 11: correlating motif scores and motif counts with Hill-fitted parameters
    • 12: calling nucleosome position with NucleoATAC
    • 13: calculate motif distance to nucleosomes
    • 14: in silico marginalization with ChromBPNet models
    • 15: ChromBPNet model performance evaluation, multinomial logistic regression models
    • 16: hit dose analysis, PWM and pileups of different hit dose sets
    • 17: overlap with ENCODE ChIP-seq data
    • 18: ChromHMM annotations
    • 19: motif pattern distribution in the genome
  • 03-rna
    • 00: snakemake pipeline to preprocess RNA data with kallisto
    • 01: PCA, differential analysis
    • 02: plot average TPMs for overexpressed TFs
    • 03: GO term enrichment for differential gene sets
  • 04-chrombpnet
    • snakemake pipeline to prepare input regions, train ChromBPNet models, interpret models, discovery motifs, and identify hit instances
  • 05-bravo
    • scripts and device configuration file for running RoboATAC on an Agilent Bravo liquid handling robot (NGS Option B layout)

Citation

If you use this data or code, please cite:

An automated ATAC-seq method reveals sequence determinants of transcription factor dose response in the open chromatin. Betty B. Liu, Masaru Shimasawa, Sidney Vermeulen, Samuel H. Kim, Nika Iremadze, Doron Lipson, Zohar Shipony, William J. Greenleaf, bioRxiv 2025.07.24.666684; doi: https://doi.org/10.1101/2025.07.24.666684