PEPATAC glossary

Information regarding universal ATAC-seq terms and definitions are available from the ENCODE consortium.

Additional information relevant to PEPATAC output and interpretation includes:

Alignment statistics

  • Raw reads: The original number of reads in the input files.
  • Trimmed reads: The number of reads remaining after trimming.
  • Trim loss rate: The percentage of total reads that was trimmed. i.e.
  • Mapped reads: The number of reads mapped to primary genome.
  • QC filtered reads: The number of reads removed due to poor MAPQ values (i.e. <10).
  • Aligned reads: (Mapped_reads - QC_filtered_reads)
  • Alignment rate: The percentage of trimmed reads that mapped to the primary genome. i.e. . In this case, trimmed reads represent the maximum number of reads that even have the potential to be mapped.
  • Total efficiency - The percentage of raw reads that mapped to the primary genome. i.e. . Here we're looking at even reads that necessitated trimming, which gives an idea of how well your sample preparation was if, for example, a large number of reads had required trimming. If the efficiency is very poor but you had a high alignment rate, it would suggest an issue with sample prep because so many reads were trimmed.
  • Unmapped reads: The number of trimmed reads that remains unmapped following prealignment and primary alignment. i.e. Trimmed_reads - Aligned_reads(prealignments) - Mapped_reads
  • Duplicate reads: The number of duplicate reads removed from the mapped reads.
  • Dedup aligned reads: The number of aligned reads following duplicate removal. i.e. Aligned_reads - Duplicate_reads
  • Dedup alignment rate: The number of deduplicated, aligned reads out of the number of trimmed reads. i.e.
  • Dedup total efficiency: The number of deduplicated, aligned reads out of the number of raw reads. i.e.

Peak calling output

The pipeline's default peak caller is MACS2. More detailed descriptions regarding MACS2 output may be discovered in the MACS2 repository.

The following files are included in default PEPATAC analyses:

  • <sample_name>_peaks.xls: An XLS formatted file containing call peak information with 1-based coordinates.
  • <sample_name>_peaks.narrowPeak: A BED6+4 format file containing peak locations, peak summits, p-values, and q-values.
  • <sample_name>_summits.bed: A BED format file containing the peak summit locations for each peak. Useful for finding motifs at these sites.
  • <sample_name>_peaks_coverage.bed: A BED format file containing the number of overlapping reads in each peak.
    Column format:
    1. chromosome name
    2. start position of peak
    3. end position of peak
    4. read count
  • <sample_name>_peaks.bigBed: A bigNarrowPeak (bigBed) formatted version of the narrowPeak file produced by MACS2. Check out the bigNarrowPeak track format page for more information.

QC output

  • <sample_name>_bamQC.tsv: A tabular file containing alignment statistics and library complexity measures.
    • total read pairs: The total number of reads (singular or paired).
    • distinct read pairs: The number of genomic locations to which some read maps uniquely.
    • one read pairs: The number of genomic locations with exactly one read mapped uniquely.
    • two read pairs: The number of genomic locations where two reads map uniquely.
    • duplicate rate: The rate of duplicate reads. Will only be >0 if this script used outside of the pipeline on a BAM file with duplicates marked.
    • mitochondrial reads: The number of mitochondrial reads. If mitochondrial prealignments are used, this should be zero or close to zero. These reads are removed prior to downstream analysis.
    • mitochondrial rate: The rate of mitochondrial aligning reads.
    • NRF (non-redundant fraction): The fraction of non-redundant mapped reads (i.e. one read pairs / total read pairs).
    • PBC1 (PCR Bottleneck Coefficient 1): one read pairs / distinct read pairs.
    • PBC2 (PCR Bottleneck Coefficient 2): one read pairs / two read pairs.
  • <sample_name>_peaks_chr_dist.[png/pdf]: Plot(s) of the chromosomal distribution of called peaks.
  • <sample_name>_<feature>_coverage.bed: A BED format file containing the number of overlapping reads in each <feature>.
  • <sample_name>_frif.[png/pdf]: Plot(s) of the fractions of reads in each <feature>. The cumulative sum of reads in a feature is plotted against the log of the number of bases covered by that feature. Therefore, it takes fewer total bases covered to visualize an enriched feature compared to others. Because peaks are called specifically on the data (i.e. enriched specifically for the data), the cumulative fraction of reads in peaks should generally be the first (left-most) curve. The number following the feature name is the fraction of reads in that feature. Features may be customized by specifying the --anno-name argument to the pipeline along with the corresponding BED file containing defined features.
  • <sample_name>_fragLen.txt: A text file containing a single column of all fragment lengths.
  • <sample_name>_fragCount.txt: A tabular file containing the number of read fragments at each fragment length.
  • <sample_name>_fragLenDistribution.txt: A tabular file containing fragment length distribution summary statistics.
  • <sample_name>_fragLenDistribution.[png/pdf]: Fragment length distribution plot(s). Normally, you should observe a well-defined peak somewhere <100-bp representing nucleosome-free regions, a second peak around 200-bp representing mono-nucleosomes, then sequentially smaller peaks representing multiple nucleosomes.
  • <sample_name>_peaks_partition_dist.[png/pdf]: Plot(s) of the number of peaks present in each defined genomic partition. Defined partitions include: 3' UTR, 5' UTR, Exonic, Intronic, Intergenic, Promoter, and Promoter Flanking Regions.
  • <sample_name>_peaks_TSS_dist.[png/pdf]: Plot(s) of the distribution of reads relative to TSS's.
  • <sample_name>_TssEnrichment.txt: A text file containing a single column of read counts at each position flanking transcription start sites (-2000 - TSS - +2000).
  • <sample_name>_TssEnrichment.[png/pdf]: Global TSS enrichment plot(s). Illustrates the aggregate enrichment around all transcription start sites.

Signal tracks

The tool to produce these tracks,, is included in the PEPATAC repository and may be called indpendently with parameters like the --smooth-length further customized.

  • nucleotide-resolution ("exact cut") signal: The nucleotide-resolution (or "exact cut") signal is a signal track (bigWig) marking exact locations where the transposition event occurred. The Tn5 transpoase acts on and duplicates a 9-bp region, and the event location is defined as the center of this 9-bp window. To produce this signal, read positions on positive strands are shifted +4 with minus strand positions shifted -5.
  • smooth signal: The smooth signal track is produced alongside the nucleotide-resolution track. Instead of marking exact bases, the smooth track marks reads positioned +/- 25-bp around the "exact cut" location, to yield a 50-bp window centered on this position.