Choice of transcript version has changed (NM_001017995.2 uses UTA 20241220)

[lhomas]

Investigate transcript cds change in UTA 20241220

Investigate cds shift by 1 base, this could possibly be an off by 1 error? Could be us, could be UTA

---

NM_001017995.2

Old transcript:

{"id":"NM_001017995.2",
"cdot":"0.2.24",
"hgnc":"29242",
"biotype":["mRNA"],
"protein":"NP_001017995.1",
"gene_name":"SH3PXD2B",
"stop_codon":2907,
"start_codon":171,
"genome_builds":{"GRCh37":{"url":"http://ftp.ncbi.nlm.nih.gov/genomes/archive/old_refseq/Homo_sapiens/ARCHIVE/ANNOTATION_RELEASE.105/GFF/ref_GRCh37.p13_top_level.gff3.gz",
"exons":[[171760502,171766920,12,1360,7777,null],[171773139,171773265,11,1234,1359,null],[171774286,171774336,10,1184,1233,null],[171777366,171777593,9,957,1183,null],[171780891,171781009,8,839,956,null],[171785776,171785881,7,734,838,null],[171789738,171789873,6,599,733,null],[171800793,171800819,5,573,598,null],[171809039,171809131,4,481,572,null],[171821566,171821643,3,404,480,null],[171833280,171833356,2,328,403,null],[171849419,171849500,1,247,327,null],[171881281,171881527,0,1,246,null]],
"contig":"NC_000005.9",
"strand":"-",
"cds_end":171881356,
"cds_start":171765372}}}

---

New in cdot 0.2.28 uses postgresql://uta.biocommons.org/uta_20241220

{"id":"NM_001017995.2",
"cdot":"0.2.28",
"protein":"NP_001017995.1",
"gene_name":"SH3PXD2B",
"stop_codon":2907,
"start_codon":171,
"genome_builds":{"GRCh37":{"url":"postgresql://uta.biocommons.org/uta_20241220",
"exons":[[171760502,171766920,12,1360,7777,null],[171773139,171773265,11,1234,1359,null],[171774286,171774336,10,1184,1233,null],[171777366,171777593,9,957,1183,null],[171780891,171781009,8,839,956,null],[171785776,171785881,7,734,838,null],[171789738,171789873,6,599,733,null],[171800793,171800819,5,573,598,null],[171809039,171809131,4,481,572,null],[171821566,171821643,3,404,480,null],[171833280,171833356,2,328,403,null],[171849419,171849500,1,247,327,null],[171881281,171881527,0,1,246,null]],
"contig":"NC_000005.9",
"strand":"-",
"cds_end":171881357,
"cds_start":171765373}}}

[davmlaw]

Change is coding start/end shifted back by 1 base (no gaps before/after)

cdot matches the UTA transcripts through resolution process. This is purely a difference in transcript alignment with UTA and not a cdot conversion issue.

NM_001017995.2:c.1A>T

In [8]: var_c = hp.parse_hgvs_variant("NM_001017995.2:c.1A>T")
In [9]: str(c_to_g(var_c))
Out[9]: 'NC_000005.10:g.172454352T>A'

Using cdot gives: NC_000005.10:g.172454352T>A - MATCH

NM_001017995.2(SH3PXD2B):c.1188+1773_2733+6592del

In [3]: var_c = hp.parse_hgvs_variant("NM_001017995.2(SH3PXD2B):c.1188+1773_2733+6592del")
In [4]: str(c_to_g(var_c))
Out[4]: 'NC_000005.10:g.172331780_172344363del'

Using VariantGrid/Shariant - NC_000005.10:g.172331780_172344363del - MATCH

[EmmaTudini]

@davmlaw Document transcript choice order/priority - e.g. UTA first then GTFs, etc.
Write text file or put in cdot json gzip
Raise new issue to investigate this as a general phenomenon to get a benchmark. Would be interesting/great paragraph in cdot paper. What is change? - is change to do with genome patch versions or aligner parameters, etc