Add variant/sample summary statistic methods

[quansight-bot]

Issue metadata

• url: https://github.com/pystatgen/sgkit/issues/162
• state: closed
• milestone: No milestone assigned
• created by: Quansight Bot @quansight-bot
• labels:
• assigned:

Issue description

Issue metadata

• url: https://github.com/pystatgen/sgkit/issues/162
• state: closed
• milestone: No milestone assigned
• created by: Quansight Bot [@quansight-bot](https://github.com/quansight-bot)
• labels:
• assigned:

Issue description

Issue metadata

• url: https://github.com/pystatgen/sgkit/issues/162
• state: open
• milestone: No milestone assigned
• created by: Quansight Bot [[@quansight-bot](https://github.com/quansight-bot)](https://github.com/quansight-bot)
• labels:
• assigned:

Issue description

Issue metadata

• url: https://github.com/pystatgen/sgkit/issues/29
• state: open
• milestone: No milestone assigned
• created by: Eric Czech [[[@eric-czech](https://github.com/eric-czech)](https://github.com/eric-czech)](https://github.com/eric-czech)
• labels:
  • core operations
• assigned:
  • Tom White [[[@tomwhite](https://github.com/tomwhite)](https://github.com/tomwhite)](https://github.com/tomwhite)

Issue description

At a minimum, we need to be able to calculate these axis-wise aggregations (for GWAS QC):

• Variants (across samples)
  • genotype counts
  • allele counts and frequency
  • call count and rate
  • hwe p value (https://github.com/pystatgen/sgkit/issues/28)
• Samples (across variants)
  • call count and rate
  • genotype counts

I would propose that we start by making methods for each that take a single Dataset and return Dataset instances (it will be easier to define the frequencies/rates when the counts are defined in the same functions).

This should consider whatever the solution to https://github.com/pystatgen/sgkit/issues/3 ends up being.