[alejandrogzi]

Why

11_bbs/ did not match 10_final/:

• Wrong transcripts. BEDTOBIGBED_STRONG_RTS, BEDTOBIGBED_WEAK_RTS and BEDTOBIGBED_ARTIFACTS converted what each filter kept (STRIP_*.out.hq), not what it removed. On HLangAng1 liver, HLangAng1_no_retentions.bb held 212,189 transcripts, while 10_final/strong_rts holds 11,960.
• Unclear names. Each BigBed took its channel id: HLangAng1_flnc@HLangAng1_flnc.discard.bb held the truncations, HLangAng1.discard.bb the retentions and HLangAng1.bb the fusions.
• No NMD track. The NMD-positive transcripts (10_final/nmd) had no BigBed.

Changes

• subworkflows/polish/main.nf: the strong-RT, weak-RT and artifact tracks sort and convert STRIP_*.out.discard. POLISH emits the sorted NMD bed (SORT_BED_NMD.out.sorted) instead of the unsorted ISOTOOLS_NMD.out.nmd; nothing else consumed that emit.
• main.nf: BEDTOBIGBED_NMD in FULL_RUN, FROM_POLISHING and FROM_ALIGNMENT.
• nextflow.config: ext.prefix = "${params.prefix ?: meta.id}.<class>" for each BigBed process, and BEDTOBIGBED_NMD published to 11_bbs. Version 0.1.18.
• README.md (the 11_bbs tree) and Changelog.md (0.1.18).

Class	Before	After
Polished transcripts	<prefix>_flnc.bb	<prefix>.flnc.bb
Truncations	<prefix>_flnc@<prefix>_flnc.discard.bb	<prefix>.truncations.bb
Intron retentions	<prefix>.discard.bb	<prefix>.retentions.bb
Strong read-throughs	<prefix>_no_retentions.bb (kept set)	<prefix>.strong_rts.bb
Weak read-throughs	<prefix>_no_retentions_no_strong_rts.bb (kept set)	<prefix>.weak_rts.bb
Artifacts	<prefix>_no_retentions_no_strong_rts_no_weak_rts.bb (kept set)	<prefix>.artifacts.bb
NMD	none	<prefix>.nmd.bb
Fusions	<prefix>.bb	<prefix>.fusions.bb
Scraps	<prefix>.scraps.bb	<prefix>.scraps.bb

fusions.free.bed, the fusion-free set, still gets no BigBed. A class with no transcripts still writes no BigBed, as before.

Testing

• nextflow config -flat for test, test-sb, test-tm, test-polish, test-rustar and test-bqc-star.
• -profile test-polish,docker --skip_bb_conversion false --xorf_call_orfs false completes and writes test.flnc.bb.
  • The test data has no read-throughs, artifacts, NMD or fusions.
  • The CI profiles set skip_bb_conversion = true, so CI does not build BigBeds.
• Real data: --from polish on HLangAng1 blood, chromosome SUPER_19 (4,186 metassembly transcripts, xORF on, production parameters).
  • The run completed in 1 h, and 11_bbs/ holds exactly the nine files below.
  • For each BigBed, bigtools bigbedtobed returns exactly the rows of its 10_final BED.
  • The counts equal the production run's on SUPER_19.

BigBed	Items	10_final rows	Same rows
HLangAng1.flnc.bb	662	662	yes
HLangAng1.truncations.bb	93	93	yes
HLangAng1.retentions.bb	418	418	yes
HLangAng1.strong_rts.bb	241	241	yes
HLangAng1.weak_rts.bb	8	8	yes
HLangAng1.artifacts.bb	245	245	yes
HLangAng1.nmd.bb	102	102	yes
HLangAng1.fusions.bb	139	139	yes
HLangAng1.scraps.bb	912	912	yes