HOMINID pipeline

HOMINID pipeline

Here are the steps in the HOMINID analysis pipeline:

1. Run hominid, which performs Lasso regression, regressing microbial abundances versus host SNP alleles.

   • Software: hominid, run with no_permutation.
   • Input files:
     • OTU or taxon table
     • host SNPs in a modified VCF format
   • Output metric of correlation: coefficient of determination, R².

2. Evaluate the significance of R² and calculate q-values via permutation testing: Run hominid with permuted sample IDs.

   • Software: hominid, run with uniform_permutation or group_permutation.

3. For SNPs with significant correlation with microbial abundances, determine the associated taxa (or other covariates) via stability selection:

   • Software hominid_stability_selection

4. Once the associated taxa are selected, it's helpful to combine the data on taxon abundances with host SNP alleles (e.g., to make plots).

   • Software hominid_sort_results