Liftover hg19 -> hg38 - Githubissues

mariacuria commented 1 week ago

Secondary issue.

mariacuria commented 3 days ago

Part of #16

mariacuria commented 2 days ago

@jeet-vora I have over 28 million GRCh37 records. UCSC LiftOver command line tool was able to map all but ~14K of them. The bulk of unmapped positions are records on chr24 which is presumably chrY (~8K) and mitochondrial DNA positions (~4K). My thoughts:

I will reach out to cBioPortal to confirm that chr24 == chrY.
Ditch mitochondrial DNA because it's not part of BioMuta.
The rest can be run through ENSEMBL Assembly Converter (the one Ned used). However, it has an issue in that the output just contains new positions and says nothing about unmapped positions. So, in the end it will be hard to tell which GRCh37 record corresponds to which GRCh38 record.

Do you have advice on how to deal with it?

mariacuria commented 1 day ago

https://crossmap.sourceforge.net/ Command-line tool and chain files that ENSEMBL uses (I'm going to use only the chain file to convert positions that UCSC tool was unable to map).

GW-HIVE / biomuta-old

Liftover hg19 -> hg38 #17