unexpected .bed file size

[olatzu]

Dear creators,

Thank you very much for your amazing tool!

I am having a bit of a trouble in the last bit. This is the error:

PLINK v2.00a3.3 64-bit (3 Jun 2022) Options in effect: --bfile vzs plink_genome_test1_ALL --memory 8192 --out plink_genome_test1_ALL_additive_0 --score plink-genome-test1_ALL_additive_0.scorefile.gz zs header-read cols=+scoresums,+denom,-fid no-mean-imputation --seed 31 --threads 2

Hostname: ce5ff77d6f31 Working directory: /Users/olatzmompeo/Desktop/NETFLOW/work/44/b47dd47c1c6fee83c29377d8aecff8 Start time: Tue Feb 21 15:36:20 2023

15986 MiB RAM detected; reserving 8192 MiB for main workspace. Using up to 2 compute threads. 1 sample (1 female, 0 males; 1 founder) loaded from plink_genome_test1_ALL.fam. 963041 variants loaded from plink_genome_test1_ALL.bim.zst. Error: Unexpected PLINK 1 .bed file size (expected 963044 bytes).

I have checked and the --input bed file and has 963044 bytes, and the plink_genome_test1_ALL.bed has 65B

Is there any command that can help with this? My original one is:

./nextflow run pgscatalog/pgsc_calc \ -profile docker \ --min_overlap 0.2 --platform arm64 --input PGS_input_test.csv --target_build GRCh37 \ --pgs_id PGS001927

Thanks a lot in advance for your time

[nebfield]

This is a strange error, because the workflow doesn't modify the bed file, except for making a copy during the PLINK2_RELABELBIM process.

You could try deleting the work directory and re-running the workflow, in case there was a temporary problem:

$ rm -r /Users/olatzmompeo/Desktop/NETFLOW/work/
$ nextflow run pgscatalog/pgsc_calc -profile docker --min_overlap 0.2 --platform arm64 --input PGS_input_test.csv --target_build GRCh37 --pgs_id PGS001927

Can I double check:

• What's the name of the original input bed file?
• What process did you follow to create the bed file? Did you convert from a VCF?

[olatzu]

Dear Nebfield,

It work perfectly! I just had to remove the duplicates in plink transforming the 23andMe files!

Thank you very very much for your time

[nebfield]

Great 🥳

I'd recommend looking at your --min_overlap parameter again, which is set quite low.

I can't give any specific analysis advice but if a scoring file (--pgs_id) requires adjusting --min_overlap to work, then I wouldn't be confident in the accuracy of the final calculated scores. I explained this a little bit more previously. You might find it's a good idea to impute your input target genomes.

I'm closing this issue because the reported problem is resolved, but please feel free to ask more questions if anything is unclear.

[olatzu]

Dear Nebfield,

Absolutely. the min_overlap is too low! Working on the imputation now.

Thanks a lot :)