I am currently doing a project regarding gene-based analysis and gene-set analysis given a certain GWAS dataset, I only know 2 well known gene-based analysis tool which is LDSC and MAGMA. As of now, I'm familiar with how MAGMA works, but not really sure about LDSC, I did read LDSC Wiki briefly, and I am leaning towards MAGMA atm, but it's probably just because I'm more familiar with MAGMA. So, which one do you think is more robust? or at least what is the pros and cons of each software? Thank you in advance.
•
written
by
andreas131298 •
For gene based and gene set based analyses, MAGMA tends to be more powerful (paper here, though MAGMA's authors are also author for this paper) and are more user-friendly.
I don't think you can run gene based analysis using LD score, and it is sometime challenging to run the pathway base analysis due to various problem (e.g. singular matrix etc). So unless there are any particular reason for you needing results from LDSC, I'd suggest just stick with MAGMA
I see, thanks a lot for the response. I saw some papers utilizing MAGMA and LDSC to detect abberant cell types for a particular phenotype, for example this paper and also, a package that tries to do cell type specific analysis has the MAGMA version and LDSC version as you find the github page here
Problem with cell type specific analysis is that they are rather fuzzy and can be very sensitive to parameter choice. There's a paper by the MAGMA group complaining the analysis procedure done by Bryois and Skene which you can find here: https://www.nature.com/articles/s41467-019-11181-1
Log in to answer this question.
