Output specification for ``atac.wdl``
=====================================

.. toctree::
   :maxdepth: 2
   :caption: Contents:

All output filenames keep prefixes from corresponding input filenames. For example. If you have started from ``REP1.fastq.gz`` and ``REP2.fastq.gz`` then corresponding alignment log for each replicate has a filename of ``REP1.flagstat.qc`` and ``REP2.flagstat.qc``, respectively.

Some summarizing QC files do not have any prefix. Find ``qc.json`` and ``qc.html`` for final QC and HTML report.

#. ``DNANexus``: If you choose to use ``dxWDL`` and run pipelines on DNANexus platform, then output will be stored on the specified output directory without any subdirectories.

#. ``Cromwell``: Otherwise ``Cromwell`` will store outputs for each task under ``cromwell-executions/[WORKFLOW_ID]/call-[TASK_NAME]/shard-[IDX]``. For all tasks except ``idr`` and ``overlap``, ``[IDX]`` means a zero-based index for each replicate but for tasks ``idr`` and ``overlap`` it stands for a zero-based index for all possible pair of replicates. For example, you have 3 replicates and all possible combination of two replicates are ``[(rep1,rep2), (rep1,rep3), (rep2,rep3)]``. Therefore, ``call-idr/shard-2`` should be an output directory for the pair of replicate 2 and 3.

For more details, refer to the file table section in an HTML report generated by the pipeline. Files marked as (E) are outputs to be uploaded during ENCODE accession.

.. csv-table:: 
    :header: Task name, File, Description
   
    trim_adapter, \*.trim.fastq.gz, adapter-trimmed FASTQ
    trim_adapter, merge_fastqs_R?_*.fastq.gz, Merged and adapter-trimmed FASTQ
    bowtie2, \*.bam, Raw BAM
    bowtie2, \*.bai, BAI for Raw BAM
    bowtie2, \*.align.log, Bowtie2 log for mapping
    bowtie2, \*.flagstat.qc, Samtools flagstat log for raw BAM
    filter, \*.nodup.bam, Filtered/deduped BAM
    filter, \*.nodup.flagstat.qc, Samtools flagstat log for filtered/deduped BAM
    filter, \*.dup.qc, Picard/sambamba markdup log
    filter, \*.pbc.qc, PBC QC log
    bam2ta, \*.tagAlign.gz, TAG-ALIGN generated from filtered BAM
    bam2ta, \*.N.tagAlign.gz, Subsampled (N reads) TAG-ALIGN generated from filtered BAM
    bam2ta, \*.tn5.tagAlign.gz, TN5-shifted TAG-ALIGN
    spr, \*.pr1.tagAlign.gz, 1st pseudo-replicated TAG-ALIGN
    spr, \*.pr2.tagAlign.gz, 2nd pseudo-replicated TAG-ALIGN
    pool_ta, \*.tagAlign.gz, Pooled TAG-ALIGN from all replciates
    xcor, \*.cc.plot.pdf, Cross-correlation plot PDF
    xcor, \*.cc.plot.png, Cross-correlation plot PNG
    xcor, \*.cc.qc, Cross-correlation analysis score log
    xcor, \*.cc.fraglen.txt, Estimated fragment length
    macs2, \*.narrowPeak.gz, NARROWPEAK
    macs2, \*.bfilt.narrowPeak.gz, Blacklist-filtered NARROWPEAK
    macs2, \*.pval.signal.bigwig, p-val signal BIGWIG
    macs2, \*.fc.signal.bigwig, fold enrichment signal BIGWIG
    macs2, \*.frip.qc, Fraction of read (TAG-ALIGN) in peaks (NARROWPEAK)
    idr, \*.*Peak.gz, IDR NARROWPEAK
    idr, \*.bfilt.*Peak.gz, Blacklist-filtered IDR NARROWPEAK
    idr, \*.txt.png, IDR plot PNG
    idr, \*.txt.gz, Unthresholded IDR output
    idr, \*.log, IDR STDOUT log
    idr, \*.frip.qc, Fraction of read (TAG-ALIGN) in peaks (IDR NARROWPEAK)
    overlap, \*.*Peak.gz, Overlapping NARROWPEAK
    overlap, \*.bfilt.*Peak.gz, Blacklist-filtered overlapping NARROWPEAK
    overlap, \*.frip.qc, Fraction of read (TAG-ALIGN) in peaks (overlapping NARROWPEAK)
    reproducibility, \*.reproducibility.qc, Reproducibililty QC log
    reproducibility, optimal_peak.gz, Optimal final peak file.
    reproducibility, conservative_peak.gz, Conservative final peak file.
    qc_report, qc.html, Final HTML QC report
    qc_report, qc.json, Final QC JSON