It adds (or makes available) new options to the user:
number of participants to simulate
extra optional hapgen2 options (mutation rate, effective population size)
It adds the ability to download ready-made leg files from S3
It adds the ability to simulate GWAS summary statistics using GTCA
Details of changes:
Added --num_participants as a param (which is then used in the simulate_gen_and_sample (haggen2) process so that user can decide how many participants to simulate (this was previously hardcoded)
Re-made the *leg files so that they now have only bi-allelic SNPs:
The *leg files were then zipped and uploaded to S3
This removal of bi-allelic SNPs is a one-off step so the pipeline does not automate this. However, all instructions to make the files are in /assets/legend_for_hapgen2/README.md. The script make-leg-files.sh is also there.
Added a download_leg_files process to download the files from S3 and unzip them
Made a new channel downloaded_leg_files_ch to handle these changes
Added a process called simulate_gwas_sum_stats that takes in PLINK files and simulate GWAS summary statistics. The user must specify --gwas_cases and --gwas_controls.
Updated documentation: README.md and docs/usage.md
Comments
Despite being uploaded to S3, I have left the *leg files, the *leg.gz (along with the README.md, the VCF and the make-leg.sh script) in assets/legend_for-hapgen2.
The extra flags for GCTA (but not the mandatory ones) are not working properly. It would seem that GTCA does not handle newlines in the same way as hapgen2 (where options can all be added together) or METAL (where padding + a \n solved the issue). I have added a card to address this: https://github.com/lifebit-ai/simulate/issues/12
mmeier93
commented
4 years ago
This PR does the following:
It adds (or makes available) new options to the user:
It adds the ability to download ready-made leg files from S3
It adds the ability to simulate GWAS summary statistics using
GTCADetails of changes:
Added
--num_participantsas a param (which is then used in thesimulate_gen_and_sample(haggen2) process so that user can decide how many participants to simulate (this was previously hardcoded)Re-made the
*legfiles so that they now have only bi-allelic SNPs:/assets/legend_for_hapgen2/README.md. The scriptmake-leg-files.shis also there.download_leg_filesprocess to download the files from S3 and unzip themdownloaded_leg_files_chto handle these changesAdded a process called
simulate_gwas_sum_statsthat takes inPLINKfiles and simulate GWAS summary statistics. The user must specify--gwas_casesand--gwas_controls.Updated documentation:
README.mdanddocs/usage.mdComments
Despite being uploaded to S3, I have left the
*legfiles, the*leg.gz(along with theREADME.md, the VCF and themake-leg.shscript) inassets/legend_for-hapgen2.The extra flags for
GCTA(but not the mandatory ones) are not working properly. It would seem thatGTCAdoes not handle newlines in the same way ashapgen2(where options can all be added together) orMETAL(where padding + a\nsolved the issue). I have added a card to address this: https://github.com/lifebit-ai/simulate/issues/12Test on CloudOS:
Current test failing (error 141).