sumstats.py mat output .mat files show zero overlapping SNPs in downstream analysis, fixed by reformatting CSV with write.table

[YaxinL]

I followed the recommended data preparation steps:
txt file → txt.gz file → csv file (with added Z-score) → mat file.

I ran: python sumstats.py mat --sumstats trait.csv --ref 9545380.ref --out trait.mat

The log showed:6468852 SNPs matched with reference file
Writing .mat file...

The results were similar for both traits.

However, when running the conjunctional FDR analysis in MATLAB:matlab -nodisplay -nosplash < runme.m
It reported 0 overlapping variants between the two .mat files, which is unexpected.

I read the original txt file in R and rewrote the formatted file using:write.table(trait, 'trait.csv', sep = '\t', quote = FALSE, row.names = FALSE)

After re-running:python sumstats.py mat --sumstats trait.csv --ref 9545380.ref --out trait.mat

The conjunctional FDR MATLAB code successfully detected overlapping SNPs and worked as expected.