The genotypes cache
What the cache is helpful for ✅
Exploratory data analysis
- If you aren't sure what scores you plan to calculate before you start then using the cache can save both time and energy 🌳
- In this case you might be exploring your data, running multiple PGS scorefiles on the same target genome files iteratively
- The cache improves performance by avoiding redundant work on variants shared across scorefiles. Many scoring files include overlapping variant sets (e.g. HapMap3 variants), even when their effect weights differ.
- With caching enabled, the workflow can skip redundant index queries and parsing previously seen variants
What the cache doesn't help ❌
Calculating many scores in parallel
- If you know before you start that you want to calculate many scores it's always fastest to run the workflow once, specifying multiple scores at runtime
- This is because scores are calculated in parallel automatically
- Calculating one score many times will be much slower than setting multiple scoring files once
Calculating a score on non-overlapping target genomes
- If you use case is to calculate one PGS on many different target genomes which don't share samples, the cache will:
- not provide any speedup
- waste storage space on your computer
- In this case it can be better to set
--publish_cache false
Loading process sequence diagram
The cache is then used during the PGSC_CALC SCORE process.
How to use the genotype cache