Precise modulation of transcription factor levels reveals drivers of dosage sensitivity

Processed data and code for "Precise modulation of transcription factor levels reveals drivers of dosage sensitivity," Naqvi et al 2022.

Count/expression data

• all.sub.150bpclust.greater2.500bp.merge.ATAC.counts.fulldep.3h.24h.txt.gz - ATAC-seq counts from all samples (SOX9 titration and depletion) over all reproducible ATAC-seq peak regions
• all.sub.150bpclust.greater2.500bp.merge.k27ac.txt.gz - H3K27ac ChIP-seq counts from SOX9 depletion samples over all reproducible peak regions
• rna.salmon.7rep.txi.counts.txt.gz - RNA-seq counts from SOX9 titration samples 
• rna.salmon.7rep.txi.abundance.txt.gz - RNA-seq TPM values from SOX9 titration samples 
• slam.tcreadcount.txt.gz - SLAM-seq T-C conversion-containing read counts (representing newly transcribed mRNAs) from SOX9 depletion samples
• slam.readcount.txt.gz - SLAM-seq read counts (representing all mRNAs) from SOX9 depletion samples

Metadata

• all.protcod.gene.features.txt.gz - Features of interest for all protein-coding genes
• all.sub.150bpclust.greater2.500bp.merge.features.txt.gz - Features of interest for all reproducible peak regions
• atac_depletion_3h_24h_design.txt - design matrix for ATAC-seq SOX9 depletion samples
• atac_titration_48h_design.txt - design matrix for ATAC-seq SOX9 titration samples
• Homo_sapiens.GRCh38.cdna.all.txid2gene.id.symbol.type.txt.gz - Ensembl transcript types (for filtering to protein-coding genes in various analyses)
• k27_depletion_3h_24h_design.txt - design matrix for H3K27ac ChIP-seq SOX9 depletion samples
• rna_titration_48h_design.txt - design matrix for RNA-seq SOX9 titration samples
• slam_depletion_3h_24h_design.txt - design matrix for SLAM-seq SOX9 depletion samples

Scripts

• atac_deseq_fitmodels_bs_parallel.R - R code for fitting bootstrapped Hill equations to all SOX9-dependent REs (computationally intensive, so has been coded for parallelization over multiple cores)
  • Input: all.sub.150bpclust.greater2.500bp.merge.ATAC.counts.fulldep.3h.24h.txt.gz, atac_titration_48h_design.txt
  • Output: enh_linear_sig_aic_bsmat.txt, enh_linear_sig_aic_bsmat_enhind.txt
• atac_deseq_fitmodels.R - R code for fitting Hill equations (no bootstrap) to all SOX9-dependent REs
  • Input: all.sub.150bpclust.greater2.500bp.merge.ATAC.counts.fulldep.3h.24h.txt.gz, atac_titration_48h_design.txt
  • Output: enh_linear_sig_aic.rds
• atac_k27_depletion_deseq.R - R code for DESeq2 analysis of ATAC and H3K27ac ChIP SOX9 depletion (3h and 24h)
  • Input: all.sub.150bpclust.greater2.500bp.merge.ATAC.counts.fulldep.3h.24h.txt.gz, all.sub.150bpclust.greater2.500bp.merge.k27ac.txt.gz, atac_depletion_3h_24h_design.txt
  • Output: atac_depletion_3h_24h_deseq.txt, k27_depletion_3h_24h_deseq.txt 
• drm.R - Modified version of code from drc() package to prevent errors, install drc() with this version to avoid errors
• group_comparisons.Rmd - R code to compare computed parameters (i.e. ED50, Hill) between sets of REs/genes utilizing bootstrap information
  • Input: enh_linear_sig_aic_bsmat_enhind.txt, enh_linear_sig_aic_bsmat.txt.gz, enh_linear_sig_aic.rds, gene_linear_sig_aic_bsmat_enhind.txt, gene_linear_sig_aic_bsmat.txt.gz, gene_linear_sig_aic.rds, all.sub.150bpclust.greater2.500bp.merge.features.txt.gz, all.protcod.gene.features.txt.gz
  • Uses: summarize_bs_helper.R
• plot_re_gene_fits.Rmd - R code for plotting individual RE/gene counts and Hill/linear fits
  • Input: all.sub.150bpclust.greater2.500bp.merge.ATAC.counts.fulldep.3h.24h.txt.gz, atac_titration_48h_design.txt, rna.salmon.7rep.txi.counts.txt.gz, rna.salmon.7rep.txi.abundance.txt.gz, rna_titration_48h_design.txt
• rna_deseq_fitmodels_bs_parallel.R -  R code for fitting bootstrapped Hill equations to all SOX9-dependent genes (computationally intensive, so has been coded for parallelization over multiple cores)
  • Input: rna.salmon.7rep.txi.counts.txt.gz, rna.salmon.7rep.txi.abundance.txt.gz, rna_titration_48h_design.txt
  • Output: gene_linear_sig_aic_bsmat.txt, gene_linear_sig_aic_bsmat_enhind.txt
• rna_deseq_fitmodels.R - R code for fitting Hill equations (no bootstrap) to all SOX9-dependent genes
  • Input: rna.salmon.7rep.txi.counts.txt.gz, rna.salmon.7rep.txi.abundance.txt.gz, rna_titration_48h_design.txt
  • Output: gene_linear_sig_aic.rds
• slam_depletion_deseq.R - R code for DESeq2/sva analysis of SLAM-seq SOX9 depletion (3h and 24h)
  • Input: slam.tcreadcount.txt.gz, slam_depletion_3h_24h_design.txt
  • Output: slam_depletion_3h_24h_deseq.txt
• summarize_bs_helper.R - Helper functions for group_comparisons.Rmd

Intermediate/output files (some files are gzipped to save space, the Rscripts that output them won't gzip but they expect gzipped input when indicated)

• atac_depletion_3h_24h_deseq.txt.gz - DESeq2 output of ATAC SOX9 depletion (3h and 24h)
• enh_linear_sig_aic_bsmat_enhind.txt - index to name file for SOX9-dependent RE bootstrap output
• enh_linear_sig_aic_bsmat.txt.gz - SOX9-dependent RE bootstrap output
• enh_linear_sig_aic.rds - Parameters from Hill equation fit on all SOX9-dependent REs (no bootstrap) (RDS file)
• gene_linear_sig_aic_bsmat_enhind.txt - index to name file for SOX9-dependent gene bootstrap output
• gene_linear_sig_aic_bsmat.txt.gz - SOX9-dependent gene bootstrap output
• gene_linear_sig_aic.rds - Parameters from Hill equation fit on all SOX9-dependent gene (no bootstrap) (RDS file)
• k27_depletion_3h_24h_deseq.txt.gz - DESeq2 output of H3K27ac ChIP-seq SOX9 depletion (3h and 24h)
• slam_depletion_3h_24h_deseq.txt.gz -  DESeq2 output of SLAM-seq SOX9 depletion (3h and 24h)

chromatin_predictions.tar.gz (self-contained folder for chromatin-based predictions of gene expression change) contains:

• ABC_6conc.sh - Bash script to calculate predicted gene expression change based on ATAC-seq fold-change at each of five SOX9 concentrations (warning: creates a number of very large additional intermediate output files). Requires as input all files in this folder except for all.sub.150bpclust.greater2.500bp.merge.ABC.5Mb.power-0.7.norm.6conc.all.total.txt
•  all.sub.150bpclust.greater2.500bp.merge.ABC.5Mb.power-0.7.norm.6conc.all.total.txt - ATAC-based predicted fold-change of all genes each of five SOX9 concentrations (78, 52, 25, 8, 0, in that order), relative to 100% SOX9
• all.sub.150bpclust.greater2.500bp.merge.ATAC.DMSO.counts.txt - ATAC-seq counts over all reproducible peak regions in updepleted samples
• all.sub.150bpclust.greater2.500bp.merge.bed - bed file of all reproducible peak regions
• all.sub.150bpclust.greater2.500bp.merge.deseq.allconc.lfc.txt - DESeq2 output from ATAC SOX9 titration, comparing each lowered SOX9 concentration to 100% SOX9
• all.sub.150bpclust.greater2.500bp.merge.k27ac.dmso.counts.txt - H3K27ac ChIP-seq counts over all reproducible peak regions in updepleted samples
• hg38_refGene_TSS_collapsed.bed - collapsed TSSs for all genes
• hg38.genome - genome file