I’m looking to lift over a dataset of GWAS summary stats from build 37 to 38, including converting rsid. Downstream it is to be used for co-localisation with another dataset of build 38.
Would anyone be able to help with code for doing this please?
$\begingroup$
$\endgroup$
2
-
1$\begingroup$ Have you tried some tool or approach already? $\endgroup$– JRodrigoFCommented Nov 18, 2022 at 10:46
-
$\begingroup$ Should we assume you are working on human data? Are the SNPs in coding regions only or do you also have intronic or, even worse, intergenic? Please edit your question and give us some context. Ideally also an example of your input. $\endgroup$– terdon ♦Commented Nov 18, 2022 at 12:12
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
There are multiple tools to use for this. Might I suggest the following reads: Introduction to liftover tools, where they also refer to this usefull post BioStars post. Maybe also consider reading this paper for a word of caution: Paper.
Maybe you can also clarify you question more, if these links do not help you with what you need.
$\begingroup$
$\endgroup$
I have recently worked on a couple of tools to do exactly this. They are +munge to convert summary statistics into files following the GWAS-VCF specification and +liftover to convert a VCF from one genome build to another. This will automatically take care of REF⇆ALT switches by flipping the effect size and it will correctly handle indels, if there are any, in your summary statistics file
