Subset columns to MIxS terms (version 5)

[wdduncan]

Create a version of biosample data whose columns are MIxS 5 terms. Not all the harmonized names (e.g., 'fire') MIxS terms.

cc @cmungall @realmarcin

[cmungall]

fire is in mixs

[cmungall]

I suggest before starting this, catalog the list of fields not in mixs, we may need to make a sssom mapping

[wdduncan]

I going to subset using the non-human environmental package terms.

• air
• soil
• sediment
• water

[wdduncan]

See notebook build-non-human-samples.ipynb: https://github.com/INCATools/biosample-analysis/blob/master/src/notebooks/build-non-human-samples.ipynb

This notebook is subset to env_packages containing the strings:

• 'air'
• 'soil'
• 'sediment'
• 'plant-associated'
• 'water'

Output has been saved to target/non-human-samples.tsv.gz.

Potential enhancements to target/non-human-samples.tsv.gz:

• include packages containing 'plant' instead of 'plant-associated'.
• normalize all package names as described in #24
• replace values indicating missing data with NaNs

cc @realmarcin