Inquiry related to using S-Predixcan for doing association study.

[Ketki Joshi]

Hello All,

I am interested in using PrediXcan to do association study.

I see that we can use S-PrediXcan and utilize GWAS summary statistics.

I was trying to use the tutorial dataset to understand how to run the code.

I have a doubt:

the GWAS data present in tutorial has these values:

  SNP  A1  A2     FRQ    INFO    BETA      SE       P

   rs915677   G   A  0.0809  0.9119 -0.0034  0.1496   0.982

whereas my GWAS summary statistics has only these columns:

hm_variant_id   hm_rsid hm_chrom        hm_pos  hm_other_allele hm_effect_allele        hm_beta hm_odds_ratio   hm_ci_lower     hm_ci_upper     hm_effect_allele_frequency      hm_code chromosome      base_pair_location      variant_id      effect_allele   other_allele    beta    standard_error  p_value effect_allele_frequency ci_lower     odds_ratio      ci_upper

I don't have FRQ and INFO column in my GWAS summary statistics. Can you let me know if it is necessary to have all the columns as well as what do they represent?

I also have one doubt: I see that in the tutorial it also covers some steps like: harmonization and imputation.

Do we need to follow these steps, or can we directly use S-predixcan to run the associations?

I will be extremely grateful if you can provide some guidance on how to proceed with doing the analysis.

Thank you.