Programs for generating Atlas of Regulated Target genes of Transcription Factors
ART-TF project includes programs for bioinformatic analysis of ChIP-seq data sets and comparison with gene expression data.
- combine_peaks_TFs.pl
Syntax: combine_peaks_TFs.pl file_list.txt peaks_all.txt -TSS TSS_symbols_hg19.txt
Perl script to associate ChIP-seq peaks with genes and assemble data from many ChIP-seq experiments into a single file. File file_list.txt is input tab-delimited text file with first column containing names of ChIP-seq experiments, and second column containing file names of corresponding ChIP-seq results. Each ChIP-seq result file is a tab-delimited text file in bed format with 4 columns: chromosome, peak start position, peak end position, peak score. Chromosome is formatted as "chr1". File TSS_symbols_hg19.txt is a tab-delimited text file with 8 columns: gene symbol, chromosome (e.g., "chr1"), position of TSS, strand (0 = forward, 1 = reverse), length of transcript (mRNA), number of exons, and accession number/ID.
- generate_targets.pl
Syntax: generate_targets.pl peaks_all.txt TF_targets_human_add_far.txt -o 1 -add
Syntax: generate_targets.pl peaks_all.txt TF_targets_human_add_near.txt -o 2 -add
Perl script to generate gene sets containing target genes for various transcription factors (TFs). Input file peaks_all.txt is the output of program combine_peaks_TFs.pl. Output file TF_targets_human_add_far.txt is a tab-delimited text file formatted as a gene-set file for ExAtlas. Each gene set is described by a line where the first item is the name of a gene set, the second item is a description or other characteristic, such as gene count, then go gene symbols (from the third item to the end of the line). Attributes of genes are optional and require additional lines. In particular, here the output file includes score as the attribute. Attribute lines in the file start wth a tab, which means that the first item is empty, and the second item is the name of attibute (here it is "score"). Scores for each gene in the set are specified starting from the third item and to the end of the line. The number of score values should match the number of genes in the previous line.
Options for generate_targets.pl are: option "-o 1" means that only distal binding cites are analyzed (at distances from 0.5 Kb to 100 Kb from TSS on both sides); option "-o 2" means that only proximal binding cites are analyzed (at distances from -0.5 Kb to +0.5 Kb from TSS); if option "-o" is not used, then all binding sites are analyzed. Option "-add" means that ChIP-seq scores are added for all binding sites linked to each gene. Each ChIP-seq peak was associated with a maximum of 3 genes whose TSS was within 100 Kb from the peak center. If th program is run without option "-add", then only one binding site is considered with the maximum score. Scores of gene/peak associations were calculated as symbol quality multiplied by the binding score (ChIP-seq) and divided by the distance from the peak to TSS (Kb, capped at 1 Kb). Symbol quality was equal 1 for "weak" symbols (e.g., containing 4 digits in a row, or strings "FAM", "MIR", "MRP", and "orf") and 3 for normal symbols. Genes with association scores <20% of the maximum value (i.e., for the best matching gene) were not reported as associated with the given ChIP-seq peak.
- The next step of the analysis is finding regulated targets of TFs; it is done using ExAtlas after gene sets of target genes are uploaded. You don't need to upload them again because these files are available in ExAtlas as public resources with the following names: public-ART-TF_targets_hg19_enhancers_add and public-ART-TF_targets_hg19_promoters_add, which specify target genes based on distal binding sites and proximal binding sites, respectively. ExAtlas is a software for meta-analysis of gene expression data. Besides standard statistical analysis of gene expression (similar to NIA Array Analysis) it supports several methods for meta-analysis and generates results for all combinations of data in multi-component data sets (e.g., all gene expression profiles vs. all GO annotations). ExAtlas demo is available at http://alexei.nfshost.com/exatlas/ (Full installation of ExAtlas: http://webtools.systemsmedicine.jp/exatlas/) and the code can be accessed at https://github.com/AlexeiSharovBaltimore/ExAtlas. To enter as guest, click button "Start using ExAtlas". Then, to access ART-TF data, users need to select organism species by clicking the pull-down menu "Select organism" shown by blue arrow at the top of the screen. Please, select "Human (Homo sapiens)".
To identify regulated targets of TFs, locate section "Select Data Files" and click on the pull-down menu in the first line (gene expression profiles). Select data "public-CREST_Human_TF_induction_ESC_20181203" which is large-scale study of the gene expression change in ES cells after induction of 510 individual TFs (Nakatake et al. 2020. Cell Reports, 31(7):107655) and click button "Open" in the same line at the left. In the next screen, locate section 4 "Other functions" and click on "Geneset analysis ...". In the next screen select geneset file "public-ART-TF_taergets_hg19_enhancers_add" and check the box "Use gene attributes". This checkbox is needed to instruct the program to use gene attributes, which in our case is the score of TF binding to enhancer(s) of each target gene. Also check another checkbox "Identify associated genes", and then click the button "Geneset analysis". The analysis may take 25-30 minutes; click button "Check your task" when it is available; also you can click in "Log file" to see the progression.
- parse_targets.pl
Syntax: parse_targets.pl TF_regulated_targets_EPFP03_add_far.txt TF_regulated_targets_EPFP03_add_near.txt target_scores_add.txt -ind indirect_list.txt
Perl script to parse output files with regulated gene targets (see step #3) downloaded from ExAtlas and named "TF_regulated_targets_EPFP03_add_far.txt" (for enhancers) and "TF_regulated_targets_EPFP03_add_near.txt" (for promoters). Option " -ind indirect_list.txt" is needed to supply the program with a file that specifies acceptable surrogate TFs so that ChIP-seq surrogate data can be used in case there is no ChIP-seq data for some induced TF. Surrogate ChIP-seq data also makes lists of regulated target genes more complete to increase their explanatoiry power. The output file target_scores_add.txt is a tab-delimited text file with 11 columns: (a) ChIP-seq data name, (b) Induced TF, (c) direct=1 or indirect=0 type of regulated target genes (indirect means using surrogate ChIP-seq data), (d) z-value for gene set enrichment of target genes with enhancer binding among upregulated genes, (e) z-value for gene set enrichment of target genes with promoter binding among upregulated genes, (f) z-value for gene set enrichment of target genes with enhancer binding among downregulated genes, (g) z-value for gene set enrichment of target genes with promoter binding among downpregulated genes, (h) number of upregulated target genes with enhancer binding, (i) number of upregulated target genes with promoter binding, (j) number of downregulated target genes with enhancer binding, (k) number of downregulated target genes with promoter binding.
- output2targets.pl
Syntax: output2targets.pl TF_regulated_targets_EPFP03_add_far.txt TF_regul_targets_add_far.txt -ind indirect_list.txt
Syntax: output2targets.pl TF_regulated_targets_EPFP03_add_near.txt TF_regul_targets_add_near.txt -ind indirect_list.txt
Perl script to extract gene sets of regulated gene targets from output files with regulated gene targets (see step #3) downloaded from ExAtlas and named "TF_regulated_targets_EPFP03_add_far.txt" (for enhancers). Option " -ind indirect_list.txt" is needed to supply the program with a file that specifies acceptable surrogate TFs. Output file TF_regul_targets_add_far.txt specifies gene symbols for regulated target genes generated from individual ChIP-seq experiments and one attribute: EPFP (expected proportion of false positives).
Geneset file format: it is a tab-delimited text file. In the simplest case, each line specifies one geneset, where first position is geneset name, second position is description, then follows a list of genesymbols (tab delimited). A geneset file can also specify attributes (e.g., gene score, p-value, or expression). Each attribute occupies a line after the line with a list of gene symbols. Such a line has emply pirst position (to distinguish it from the list of genes. The second position is the name of attribute, then follow attribute values for all genes in the geneset (tab delimited). All genesets in a file should have the same attributes.
- target_overlap_v2.pl
Syntax: target_overlap_v2.pl TF_regul_targets_add_far.txt ndata_chip.txt target_overlap_add_far.txt
Syntax: target_overlap_v2.pl TF_regul_targets_add_far.txt ndata_chip.txt TF_regul_targets_add_far_combined.txt -g
Syntax: target_overlap_v2.pl TF_regul_targets_add_near.txt ndata_chip.txt target_overlap_add_near.txt
Syntax: target_overlap_v2.pl TF_regul_targets_add_near.txt ndata_chip.txt TF_regul_targets_add_near_combined.txt -g
Perl script to count overlapping regulated gene targets for pairs of gene sets obtained from different ChIP-seq experiments with the same TF gene, and also to generate a file with combined genesets. Input file TF_regul_targets_add_far.txt is generated at step 5; it is a list of genesets of reguilated targets of TFs. Second input file ndata_chip.txt is a text tab-delimited file with three columns: TF symbol, Ndata_all, and Ndata_direct. Ndata_all is the total number of ChIP-seq data that successfully yielded some regulated targets for a TF; Ndata_direct is the number of non-surrogate ChIP-seq data (i.e., direct) that yielded some regulated targets for a TF.
If option "-g" is added at the end, then the program generates combined gene sets of regulated target genes from multiple ChIP-seq experiments with the same TF gene. For each TF, up to 4 gene sets can be generated: upregulated direct targets, downregulated direct targets, upregulated indirect targets, and downregulated indirect targets. Two attributes for target genes are generated: smallest EPFP among ChIP-seq data sets, and the number of ChIP-seq data sets that support each regulated target gene. If 3 or more ChIP-seq data sets are available for a TF, then regulated target genes suported by a signle ChIP-seq data set are not included into a combined gene set of regulated targets.
- compare_indirect_direct.pl
Syntax: compare_indirect_direct.pl target_overlap_add_far.txt freq_indirect_direct_far.txt
Perl script to compare the quality of direct and indirect regulated targets. The significance of overlap between sets of direct and indirect regulated targets for the same TF was quantified by the hyporgeometric test (z-value). The overlap between sets of direct and indirect regulated targets was generally lower than the overlap between sets of direct regulated targets identified using different ChIP-seq data for the same TF, as follows from the probability distribution of z-values.
- target_far_near_new.pl
Syntax: target_far_near_new.pl TF_regul_targets_add_far_combined.txt TF_regul_targets_add_near_combined.txt regul_targets_table_add.txt
Syntax: target_far_near_new.pl TF_regul_targets_add_far_combined.txt TF_regul_targets_add_near_combined.txt regul_all_targets_genesets_add.txt -gall -anova anova-TFs_Sep2018_all.txt -coord TF_targets_human_add_far.txt TF_targets_human_add_near.txt -indir indirect_list.txt
Perl script to generate a table of varyoius categories of regulated gene counts: upregulated and downregulated, direct and indirect, bound by TFs in enhancers or prtomoters or both. Options "-gall" and "-gdir" are used to generate a file with genesets of regulated targtes supported by all ChIP-seq data or only by non-surrogate ChIP-seq data (direct), respectively. These gene sets are characterized by the following attributes: (a) smallest EPFP, (b) average logratio (log10) of gene expression channge after induction of the TF, (c) counts of ChIP-seq data sets supporting the target gene. and (d) genome position of the best binding site of the TF (hg19) or a pair of binding sites one in enhancer and another in promoter.
- count_2fold_targets.pl
Syntax: count_2fold_targets.pl regul_all_targets_genesets_add.txt signif_Nov2018_2fold_MedEmerCag_genesets.txt regul_all_targets_2fold.txt
Perl script to generate genesets of regulated targets that also show signiuficant change of expression (>2 fold & FDR<0.05) in ES cells after induction of corresponding TFs. Both input files and the output file (regul_all_targets_2fold.txt) are gormatted as genesets for ExAtlas (as explained above).
- generate_table_targets_v2.pl
Syntax: generate_table_targets_v2.pl regul_dir_targets_genesets_add.txt regul_all_targets_genesets_add.txt output.txt
Perl script to generate a summary table of regulated target genes for all TFs. Output file is a tab-delimited table/text with 11 columns. Column headers: a) Induced transcription factor (TF) gene symbol; b) Direction of gene expression change after TF induction; c) Target gene number ordered by increasing EPFP; d) Target gene symbol; e) Log10-ratio of target gene expression change after TF induction; f) All regulated targets (including surrogate ChIP-seq): EPFP; g) All regulated targets: Number of supporting ChIP-seq data sets; h) All regulated targets: genome position (center); i) Direct regulated targets (no surrogate ChIP-seq): EPFP; j) Direct regulated targets: Number of supporting ChIP-seq data sets; k) Direct regulated targets: genome position (center).
- geneset_match.pl
Syntax: geneset_match.pl geneset1.txt geneset2.txt output.txt
Perl program to compare genesets in files geneset1.txt and geneset2.txt (file format is in #5). The output is a tab-delimited table with 4 columns: (a) geneset name, (b) n genes in file1, (c) n genes in file2, and (d) n genes that match between file1 and file2.