PR Log

This tracks all PRs made since the last Stoat Report. Ideally we should add a summary of each PR to be added to the report.

PR History

• #7: Simulation lr
• #8: Change the stoat graph subcommand to take the same binary phenotype file as stoat vcf
• #9: Switch old uncompiled debug code to new logger
• #10: Refactor the partitioning code to use the same code path for regular snarls and irregular snarls.
  • Also add a way to use assert() with the logger
• #11: Handle special case collinearity lr, Remove BH & top variant file from natif stoat vcf/graph, add verbosity BH, remove LOG_FATAL replaced by throw, Add namespace stoat in stat_test.*
• #12: Change the actual stoat lr implementation by the mathematical one, correct filtration, handle new specific table regression case (pseudo inversion needed).
• #15: Fix bug in stoat graph looking at paths we didn't include
• #16: Add a script for making the distance index with -w to try and force the top-level chain to include nodes with all sample haplotypes *This could be more flexible in choosing the nodes because if a path is fragmented it won't find any nodes, but it depends a lot on the graph and it works for now
• #17: Add more robust parsing condition for VCF file and unify input graph arg
• #18: Make sure that all paths get indexed for the path position overlay
• #19: Fix bug where BH correction overwrites a column
• #21: Write snarl path partitions to a file and multithread stoat graph
• #23: Add MAF filtering for stoat graph
• Fix bugs in finding snarl coordinates and parallelization
• #32: Big refactor
  • Using new Tables to combine genotypes and phenotypes
  • Homogenized stoat vcf and stoat graph using the same SnarlDataCollection to manage snarl information and common temporary files
  • Move the testing to a new test subcommand.
  • Reviewed regression code and implemented Firth penalized logistic regression.
  • Major cleanup of unused functions and objects, and renaming effort
  • Fix a bug when the orientation in the VCF and the prepared snarls don't match (stoat vcf)
  • New option in stoat vcf to try to resolve nested variants that might not be properly called in the top-level snarl.
  • New subcommand change-ref to change the reference paths of STOAT's output file without having to rerun everything.
  • Files are written as bgzipped TSV files by default to save disk space.
• #58: Adding new log warning argument: number of time this warning as been seen, the goal is to stop printing on terminal warning after N time (where N is 5 for now) BUT still conserve all warning in the log file. Also add information/log info concerning snarl number found, paths found, etc ...