search

lh3 / minigraph

Sequence-to-graph mapper and graph generator
https://lh3.github.io/minigraph
MIT License
410 stars 39 forks source link

--call outputs a path that doesn't exist in the graph #35

Open wwliao opened 3 years ago

wwliao commented 3 years ago

Hi Heng,

I notice a path >s48420>s48421>s48428 in the --call outputs that doesn't actually exist in the GRCh38-freeze1.gfa because there is no such link between s48421 and s48428. Here are the calls containing this path:

chr7    144186940   144296182   >s48420 >s48428 >s48421:10670:-:HG00438#2#h2tg000019l:15485853:15496541
chr7    144186940   144296182   >s48420 >s48428 >s48421:10670:-:HG01123#1#h1tg000013l:13689309:13699991
chr7    144186940   144296182   >s48420 >s48428 >s48421:10670:+:HG01243#2#h2tg000196l:341837:352526
chr7    144186940   144296182   >s48420 >s48428 >s48421:10670:+:HG02080#2#h2tg000103l:15430162:15440860
chr7    144186940   144296182   >s48420 >s48428 >s48421:10670:+:HG03492#2#h2tg000104l:354601:365286

The GRCh38-freeze1.gfa and its calls are both from ftp://ftp.dfci.harvard.edu/pub/hli/minigraph/HPP/

I'm not sure how to interpret this result.

Thanks, Wen-Wei

lh3 commented 3 years ago

This is probably a bug. You can take it as uncalled . for now.

lh3 commented 3 years ago

I can confirm it is a bug caused by inversions. It is difficult to fix it. For now, take such inconsistencies as uncalled. Cases like this should be fairly rare.

wwliao commented 3 years ago

Thank you for looking into this!