This repo contains code associated with re-running certain steps of snpArcher (for example, if we decide to remove an additional sample) and running downstream analyses. In brief this means:
This repo contains the following snakemake modules from snpArcher:
- postprocess
- ccgp
- trackhubs
On top of this, some code has been added to the ccgp module. The distinction between ccgp and algatr at the moment is nebulous and warrants revisiting sometime later, probably they can be made the same module or possibly algatr modded down to be more simplified.
New modules here include:
- algatr
- local_pca
- postprocess_variant
- Install mamba and snakemake
- Clone this repo
- Run the following to test out a small subset of data (within the repo directory):
snakemake --dir test_data/ --use-conda -c 10 --profile profiles/silverbullet_reruns/
If you want to run on a different project, you must match the file naming and structure of the test_data directory. This contains a handful of outputs from the snpArcher workflow:
- Raw vcf
- Callable sites bed file
- List of samples to remove
- The fai index
A config file is required named config.yaml which defines the parameters for the run. This is similar to the snpArcher config, but has a few custom additions.
- clean SNPs has MAF filter set of < 0.01
- setting MAF filter to < 0.05 for the mil sub sampled SNP file. But this hasnt been re run yet
- All VCFs should be filtered for F_MISSING < 0.25 (must be 75% of individuals with data)