smunger

munger for GWAS summary statistics

• PyPI: https://pypi.org/project/smunger/
• Free software: MIT

Features

• define column properties

  • required columns: CHR, BP, EA, NEA
  • optional columns: BETA, SE, P, EAF, MAF
  • Auxiliary columns: OR, OR_SE, Z
  • Data types
  • Data ranges
  • Allow missing values and default missing values

• semi-automatically header mapping

  • read first five rows and display in terminal
  • guess header map by common column names
  • manually check if the mapping is correct
  • input the right column number if it is wrong
  • check if OR, OR_SE, Z are present if BETA, SE are absent
  • save the final column map to json for further munging

• data munging

  • EA ≠ NEA
  • if EAF presents, MAF = min(EAF, 1-EAF)
  • convert OR/ORSE to BETA/SE, if BETA, SE are absent and OR, ORSE are present
  • remove duplicate SNPs with same chr-bp-sorted(EA,NEA), keep the one with lowest P
  • output: \t separated, bgzip compress, tabix index.
  • optional output: significant SNPs, munge report

  	CHR	BP	rsID	EA	NEA	EAF	MAF	BETA	SE	P	OR	OR_SE	Z
  type	int	int	str	str	str	float	float	float	float	float	float	float	float
  allow null	False	False	True	False	False	False	False	True	False	True	True	False	True
  null value								0		0.999	1		0
  range	[1，23]	(0,inf)		only contains ‘ACGT’	only contains ‘ACGT’	[0,1]	[0,0.5]	(-inf,inf)	(0, inf)	(0,1)	(0, inf)	(0, inf)	(-inf,inf)

• liftover

  • guess genome build
  • liftover

• annotate

  • annotate rsID

Credits

This package was created with Cookiecutter and the waynerv/cookiecutter-pypackage project template.