Output results as *.gz for LDSC ready format

[Al-Murphy]

@bschilder might be worth us looking into adding this functionality to your branch, it was asked for from a user. Do you know what exactly is required for this format?

[bschilder]

Been a while since i used LDSC but im sure i could figure it out. Will add to my branch (bschilder_dev) once I do.

[Al-Murphy]

Great, let me know if I can help with this? I have yet to actually run LDSC myself though!

[bschilder]

Info here: https://github.com/bulik/ldsc/wiki/Summary-Statistics-File-Format

I think we have most things by default except:

• Z
• Guaranteeing N is present (we can let users supply a single N value to fill in every row if necessary)
• Making A1/A2 correspond to effect/other alleles (instead of ref/alt like we currently do). Should we consider adding a EA (effect allele) column to MungeSumstats to keep track of this info?

That said, Kitty was able to run Open GWAS sum stats i just processed with MungeSumstats, but only after running LDSC's munge_sumstats.py https://github.com/bulik/ldsc/blob/master/munge_sumstats.py

[Al-Murphy]

Yep so Roxy said a similar thing that after running munge_sumstats.py it was fine. We should probably look at that function? My interpretation was that it just removed unnecessary columns?

[bschilder]

Exporting to LDSC format is now supported by setting format_sumstats(ldsc_format=TRUE, ...)

New functions to support this are:

check_ldsc_format

https://github.com/neurogenomics/MungeSumstats/blob/bschilder_dev/R/check_ldsc_format.R

check_zscore

https://github.com/neurogenomics/MungeSumstats/blob/bschilder_dev/R/check_zscore.R

[bschilder]

Passed force_new_z up to format_sumstats and set default to FALSE.