Commits for angie

7d50cb16e1ab201dfaead9f47d63fb3636ae5020 Wed May 2 14:41:09 2018 -0700

Replacing an assert that caused blank page with an errAbort message linking to trackDb doc. refs #21386
- src/hg/lib/trackDbCustom.c - lines changed 4, context: html, text, full: html, text

9a9e0448c012df593844f2a16c70ca4526d2cf6f Fri Apr 27 09:34:21 2018 -0700

Changes to gff3ToRefLink.pl to avoid missing occasional protAcc, pseudo, gene_biotype, gene_synonym. * Instead of accumulating tag values as we go, first parse the GFF into a tree and traverse the tree to find the right level of transcripts to output and the tags for those transcripts. Tolerate some kinds of mismatches, gloss over per-transcript-mapping differences, report other mismatches. * Use raFile prt value as a backup in case GFF is missing protein_id because only a non-coding exon(s) of a coding transcript could be mapped to the reference assembly, so no CDS records in the GFF. * Use hash-membership to select which tags to keep and which model organism external database IDs to extract from Dbxref. no redmine - email thread "ncbiRefSeq QA" in which Braney reported missing NP_.
- src/hg/utils/automation/doNcbiRefSeq.pl - lines changed 4, context: html, text, full: html, text
- src/hg/utils/automation/gff3ToRefLink.pl - lines changed 511, context: html, text, full: html, text

f87ac2bb68a10bfb3cb35f929586797d6ffe9cfd Fri May 4 17:29:01 2018 -0700

Obscure new instances of reused IDs keep cropping up in various species' GFF from NCBI, but not affecting the RefSeq transcripts. warn instead of die.
- src/hg/utils/automation/gff3ToRefLink.pl - lines changed 2, context: html, text, full: html, text