Repeated ARIC weights

[Zurodher]

Hi,

I noticed that in the PrediXcan PP ARIC weights there are repeated weight values for a given SNP. See below an example for ARIC EA weights:

gene            rsid        varID                  ref_allele eff_allele    weight

2396  ENSG00000165124 rs1341983 chr9_109883588_G_A_b38    G           A           0.009390807

2752  ENSG00000165124 rs1341983 chr9_109883588_G_A_b38    G           A           0.003035701

When looking at the weights for the same rsID (rs1341983) from the original data source (http://nilanjanchatterjeelab.org/pwas/) I noticed some discrepancies:

"/…/Plasma_Protein_weights_EA/SeqId_11109_56.wgt.RDat"

       enet       top1

0.003710933 2.056933819

Here are my questions:

1.       1. Why are there repeated values in PrediXcan ARIC weights? How does PrediXcan manage repeated weights for a given SNP?

2.      2.  Why are PrediXcan ARIC weights different from the original data source?

Many thanks in advance,

Zulema

[Hae Kyung Im]

Thank you, Zulema, for pointing that out.

Sofia: could you please investigate?

Sabrina: I believe you made these files. Do you know how we could have gotten duplicated SNPs with different weights?

Thanks

Haky

--
You received this message because you are subscribed to the Google Groups "PrediXcan/MetaXcan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to predixcanmetax...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/predixcanmetaxcan/f8b4131d-3258-45f8-9e43-fefb327064cen%40googlegroups.com.