WIP Breaking down the PCA user story

[jerowe]

@alimanfoo @dharhas @daletovar

I went through the tour of scikit allele blog post and broke it down.

This comes from the example notebook and the tour of scikit-allel blog post.

Here's an overview:

1. Looked at distributions to give us an overview of the Variant Attributes in 001-Exploratory-Statistics-Variant-Attributes
2. Looked at distributions to give us an overview of Variant Quality in 002-Exploratory-Statistics-Variant-Quality

Once we've done the initial exploratory analysis we want to start to remove variants that are of poor quality (or some other metric) and variants that don't give us any information on segrating our populations.

3. Filtered variants based on Variant Attributes and Variant Quality in 003-Filter-Variants
4. Filtered variants based on their Sample Missingingness, Sample Heterozygosity, and Variants that don't segregate our populations in 004-Exploratory-Statistics-Sample-QC.

All along the way, we have continually filtered our genotypes based on findings from the exploratory statistics and QC.

[daletovar]

Thanks for putting this together, @jerowe. Unless you have anything more that you'd like to add I think we can merge this.

[jeromekelleher]

This looks good to me, but I'd rather wait for @alimanfoo's input before merging. Is keeping this open for a while going to block people's progress?

[jerowe]

@jeromekelleher no we're fine. This is a WIP.

[jerowe]

Now that I have the Xarray format straight in my head I want to convert these to use the Genotype Call Dataset.