Missing sampels in clinical data causing wrong prevalence calculated

[sainadfensi]

Describe the issue

Some sampels is missing from the clinical data if no mutations were found. This will make prevalence looks higher than the actual number.

Command

maf = data.table::fread("xx")
maf$NCBI_Build = as.character(maf$NCBI_Build)
maf$NCBI_Build = "GRCh38"  

dat=dat[dat$column==Group,] # for my data, it should be 107 samples
maf=maf[Tumor_Sample_Barcode %in% dat$tmid,] # only 90 were found having variants 

laml = read.maf(maf = maf,
                    removeDuplicatedVariants=FALSE,
                    vc_nonSyn=c("Frame_Shift_Del", "Frame_Shift_Ins", "Splice_Site",  "Translation_Start_Site","Nonsense_Mutation", "Nonstop_Mutation", "In_Frame_Del","In_Frame_Ins", "Missense_Mutation","Non-coding"),
                    verbose = FALSE,
                    cnTable = cn_table,
                    clinicalData = dat)

After loading, although the clinicalData is set as using dat, only 90 samples:

nrow(laml@clinical.data)
[1] 90

I'm using version 2.22.0.

[Frank1219]

I also found this error.
Is there some suggestions ?

[sainadfensi]

Hi @Frank1219,
just for a temporary solution, I made some fake mutations for those missing samples. Although the mutaions will be filtered for mutational types as synonymous, their clinical information could be registerd when loading.

say the maf data is called maf, clinical data is called dat

 if(length(unique(maf$Tumor_Sample_Barcode))<nrow(dat) ){
    pseudo=data.frame(Tumor_Sample_Barcode=dat$tmid[!dat$tmid %in% maf$Tumor_Sample_Barcode]) %>% 
      mutate(Hugo_Symbol="pesudo",
             Chromosome="chr1",Start_Position=1,  End_Position=2,Variant_Classification="something",
             Variant_Type="pseudo",Reference_Allele="-",Tumor_Seq_Allele1="-", Tumor_Seq_Allele2="-",
             Entrez_Gene_Id="UNK", NCBI_Build="GRCh38"
             )
    maf=rbind(maf,pseudo)
  }

[Frank1219]

Hi @Frank1219, just for a temporary solution, I made some fake mutations for those missing samples. Although the mutaions will be filtered for mutational types as synonymous, their clinical information could be registerd when loading.

say the maf data is called maf, clinical data is called dat

 if(length(unique(maf$Tumor_Sample_Barcode))<nrow(dat) ){
    pseudo=data.frame(Tumor_Sample_Barcode=dat$tmid[!dat$tmid %in% maf$Tumor_Sample_Barcode]) %>% 
      mutate(Hugo_Symbol="pesudo",
             Chromosome="chr1",Start_Position=1,  End_Position=2,Variant_Classification="something",
             Variant_Type="pseudo",Reference_Allele="-",Tumor_Seq_Allele1="-", Tumor_Seq_Allele2="-",
             Entrez_Gene_Id="UNK", NCBI_Build="GRCh38"
             )
    maf=rbind(maf,pseudo)
  }

Yeah, that is indeed a feasible temporary solution.
Thanks a lot.

[github-actions]

This issue is stale because it has been open for 60 days with no activity.

[github-actions]

This issue was closed because it has been inactive for 14 days since being marked as stale.