Appendix D — File formats
This appendix lists the files you can give EpiNexus and the main tables it writes, with their columns. Chapter 3 explains the common genomic formats.
D.1 Files you can upload
| Kind | File endings | Notes |
|---|---|---|
| Reads | .fastq, .fq, .fastq.gz, .fq.gz |
Single-end, or paired-end as two files (read 1 and read 2) |
| Alignments | .bam, with .bai or .bam.bai |
Sorted BAM files; upload the peak files with them (EpiNexus calls peaks only from FASTQ files) |
| Regions | .bed, .narrowPeak, .broadPeak (also .gz) |
Peaks, for runs that start from BAM files |
| Signal | .bedGraph, .bdg, .bigWig, .bw |
Signal tracks |
| Methylation | .cov, .cov.gz |
Bismark coverage files |
| Sample sheets and tables | .csv, .tsv, .txt |
Sample sheets |
Sample sheet. A sample sheet has one row per sample, with the columns SampleID, fastq_r1, fastq_r2, Factor, Condition and Replicate (Chapter 4). Leave fastq_r2 empty for single-end reads. Factor is the target (for example H3K27ac), or input or IgG for a control. Three columns are optional: Tissue and CellType (free text) and TissueGroup (the reference tissue for the tissue check). When a sheet does not have these columns, EpiNexus keeps the tissues already entered for samples with the same name.
D.2 Tables EpiNexus writes
All tables are in the run’s results. All except the Regulatory Atlas files are also in the GEO submission package (Chapter 20). Coordinates are 0-based and half-open, as in BED files.
| File | One row per | Columns |
|---|---|---|
all_peaks.csv |
tested region (differential) | baseMean (mean normalised count), log2FC (shrunk), lfcSE, stat, pvalue, FDR, log2FC_mle (unshrunk), peak_id, direction, chrom, start, end, significant; for regions split into pieces (since 1.3.0) also parent_id (the region a piece belongs to) and n_peaks |
all_peaks.csv from csaw (broad marks) |
tested region | peak_id, chrom, start, end, log2FC (not shrunk), pvalue, FDR, baseMean (empty), direction, significant; no size_factors.csv and no normalised counts |
size_factors.csv |
sample | sample, size_factor |
counts_raw.tsv.gz |
region | region_id, chrom, start, end, then the read count of each sample |
counts_normalized.tsv.gz |
region | the same, with counts divided by the size factors |
<sample>_annotated.csv |
peak | chrom, start, end, name, score, peak_center, gene_id, gene_name, distance_to_tss, annotation, gene_strand |
GO_enrichment.<set>.csv |
GO term | ontology, term_id, term_name, region_hits, expected_hits, fold_enrichment, binom_p, binom_fdr (region test), genes_in_term, gene_hits, hyper_p, hyper_fdr (gene test), significant, genes |
differential_motifs.tsv |
motif and region set | set, motif_id, name, tf_class, tf_family, consensus, tp and tp_pct (regions with the motif), fp and fp_pct (control with the motif), enr_ratio, pvalue, qvalue, significant |
monalisa.tsv |
motif and fold-change bin | motif_id, name, bin, log2enr, negLog10P, negLog10Padj |
<sample>_super_enhancers.csv |
joined enhancer region | id, chrom, start, end, size, signal, rank, is_super, nearest_gene |
<sample>_links.csv |
enhancer–gene link | peak_id, chrom, start, end, element_class, is_self_promoter, gene_id, gene_name, gene_tss, distance, contact, activity, abc_score |
integration_genes.csv |
protein-coding gene | gene_id, gene, chrom, tss, strand, group; for each mark (h3k27ac_, h3k27me3_): the promoter call, its regions, gained, lost, log2fc, fdr and best region, and the nearby call with its regions, gained and lost (Chapter 14) |
qc_report.csv |
sample | reads, usable fragments, mapping, duplicates, peaks, FRiP, complexity (NRF, PBC1, PBC2), cross-correlation (NSC, RSC), assay-specific metrics, spike-in metrics, the class the sample was graded as, and the status (PASS, WARN or FAIL) with its reasons |
<set>.labels.tsv.gz (Regulatory Atlas) |
region of a set | chrom, start, end, element_class and element_id (the ENCODE element it overlaps), universal_state (the genome-wide chromatin state), active_in and measured_in (in how many reference samples the element is active, of how many measured), top_tissue |
Peak files from MACS3 follow the narrowPeak format: chromosome, start, end, name, score, strand, signal value, −log10 p-value, −log10 q-value and the summit’s position within the peak (Chapter 7).