Intermediates and Filtering

After the initial group calling the BRASS flow progressively filters the groups down via several methods.

Some of the intermediate files are available in *.intermediates.tar.gz ordered as follows:

1. .groups.gz - raw brass group output
2. .groups.filtered.bedpe
3. .groups.clean.bedpe

groups.gz

This is the raw output from the grouping algorithm:

• cols 1-4 LOW chr, strand, start, stop
• cols 5-8 HIGH chr, strand, start, stop
• N columns of read counts from each sample (see #NSAMPLES in header) ordered as per the header
• Hits repeat (. unless repeat filtering was used)
• N columns of read NAMES from each sample (see #NSAMPLES in header) ordered as per the header

Intermediate files

The *.brass.intermediates.tar.gz contains files that are useful for debugging and deep investigation.

Content as displayed via tar ztf *.brass.intermediates.tar.gz.

• ${T} = Tumour
• ${N} = Normal

Many files are extended versions of the window GC reference input bed file (tagged as WGC-fmt in table). Format as follows:

Column	Description
1	Chromosome/contig
2	start (0-based)
3	end (1-based)
4	b.p. of non N sequence in window
5	Fraction of bases GC (gc_bp / non_n_bp), NA when col-4 is 0

File listing, ordered by creation:

File	Description
intermediates/samp_stats.txt	Purity/ploidy, male/female status. Inputs provided at execution
intermediates/${T}_vs_${N}.groups.gz	Primary grouping with normal panel filtering
intermediates/${T}_vs_${N}.groups.filtered.bedpe	Groups passing basic blat and read support filtering.
intermediates/${T}.insert_size_distr	Corrected insert size distribution using samtools view -f 66 -F 3868 as filter based on chr5/5
intermediates/${T}_vs_${N}.ngscn.abs_cn.bg.gz	BedGraph version of absolute copynumber
intermediates/${T}_vs_${N}.ngscn.segments.abs_cn.bg.gz	Segmented absolute copynumber
intermediates/${N}.ngscn.bed.gz	Normal: WGC-fmt + count of properly paired reads
intermediates/${N}.ngscn.fb_reads.bed.gz	Normal: As intermediates/${N}.ngscn.bed.gz + reads on same strand and contig (foldback).
intermediates/${T}.ngscn.bed.gz	Tumour: WGC-fmt + count of properly paired reads
intermediates/${T}.ngscn.fb_reads.bed.gz	Tumour: As intermediates/${N}.ngscn.bed.gz + reads on same strand and contig (foldback).
intermediates/${T}_vs_${N}.is_fb_artefact.txt	List of event IDs considered to be fold-back artefacts (metropolis_hastings_inversions.R).
intermediates/${T}_vs_${N}.r2	filter_small_deletions_and_fb_artefacts.R
intermediates/${T}_vs_${N}.r3	Identifies groups that should be merged
intermediates/${T}_vs_${N}.r4	Corrects breackpoints using clipped reads (get_abs_bkpts_from_clipped_reads.pl)
intermediates/${T}_vs_${N}.r5[.scores]	Filter events due to microbial or viral sequences (filter_with_microbes_and_remapping.pl)
intermediates/${T}_vs_${N}.ngscn.abs_cn.bg.rg_cns.gz	Combined segmentation of tumour/normal taking into account cent/telo and purity data. (get_rg_cns.R)
intermediates/${T}_vs_${N}.r6	Adds flag where event hits a copy-number change as defined in intermediates/${T}_vs_${N}.ngscn.abs_cn.bg.rg_cns.gz
intermediates/${T}_vs_${N}.cn_filtered	Filtered version of r5, downstream input
intermediates/${T}_vs_${N}.groups.clean.bedpe	Annotates translocations, occurrences, copynumber changepoints, L v H range blat scores
intermediates/${T}_vs_${N}.inversions.pdf	Plot showing inversions, insert size distribution and bad groupings. Intended for debugging.