Why are phenotype values missing in PLINK analysis despite no NA values in the input?

Question

[comu]

Hi, I encountered an issue while performing SNP analysis using PLINK. Before proceeding with PRS calculation using LDpred2 and PRSice-2, I conducted a phenotype-based SNP analysis. Here's the situation:

I used a snp_filtered_binary file as input and two different phenotype .txt files for analysis. All NA (missing) values were already excluded from the phenotype files beforehand. However, when running the analysis, I noticed discrepancies in the number of phenotype values reported in the output:

1. For the first phenotype file, the output states:

   6441 phenotype values present after --pheno.
   

   The actual number of individuals without missing data should be 7246, meaning 805 phenotype data points are unexpectedly missing.

2. For the second phenotype file, the output states:

   3201 phenotype values present after --pheno.
   

   The actual number of individuals without missing data should be 3436, meaning 235 phenotype data points are unexpectedly missing.

Here are the options I used for this analysis:

--bfile /serotonin_snps_filtered_binary \
--autosome \
--linear \
--beta \
--ci 0.95 \
--pheno /Continuous_Clean.txt \
--out /Continuous_Clean_NoX

Could anyone help me understand why these individuals are being excluded? Thank you for your time!

[chrchang523]

Why do you think the .fam file still contains every sample in your phenotype files?