tjmeyers
commented
6 years ago Update: I ran the program using the inputs from your analysis: the example.stat.rds sumstats file, TWAS weights CMC.BRAIN.RNASEQ from the Fusion website, LDREF data = 1000G.EUR. , and I used the Hallmark gene sets collection from MSigDB (shorter than GO/KEGG). I ran your R scripts (aSPUpath2.R with dist_support.R) in R 3.4.3 and I loaded all the dependencies.
I see the same issue I was having with my own inputs: the output claims non-overlapping Z-scores and a p-value is only calculated for the first pathway. I've attached both output files here. Could you please help me figure out what's going on and how to fix it? Thanks. pbs_output.txt output.txt
I'm having a persistent issue where the R program is running but not calculating p-values for the pathways. Here's what the data and script output look like:
[script output:] WARNING : NA NA NA NA NA had no overlapping GWAS Z-scores [1] 22 ERROR : non-numeric matrix extent WARNING : NA NA NA NA NA had no overlapping GWAS Z-scores [1] 19 ERROR : non-numeric matrix extent WARNING : GTEx.[tissue]/GTEx.[tissue].[geneid].wgt.RDat [wgt input] had no overlapping GWAS Z-scores
[data output:] "pathway" "# genes" "# SNPs" "PathSPU(1)" "PathSPU(2)" "aSPUpath2" "time" "1" [Pathway] "1" "20" "0.992689196739919" "0.573579089190823" "0.818165206824672" "0.0879999999999654" "2" NA NA NA NA NA NA NA "3" NA NA NA NA NA NA NA [etc for all remaining pathways]
The non-overlap issue doesn't make sense because my summary stats have already been imputed to 1000 Genomes.