Cost of the SAPPHIRE Phase Polishing pipeline on the 200k UK Biobank release.

The cost for all three steps of the phase polishing pipeline for chromosomes 1–22 of the UK Biobank 200k release. The final number of polished genotypes (GTs) and rephased GTs is given for all chromosomes. Notes: Five-fold increases in cost are for jobs that were interrupted and had to be relaunched with higher priority. For example, extraction on chromosome 6 and 7 almost take the same time on the same machine but the job of chromosome 6 was interrupted and relaunched at a higher priority (and cost). Chromosome 3 had the phase calling jobs split into batches of 10,000 samples as a test, which showed that it was better to split big chromosomes in more batches of smaller sample size. This has two advantages, first, the wall clock time is reduced and second, the cost is reduced because with the large sample size, the jobs had to run for a long time and were interrupted and had to be relaunched with high priority which increased the cost.
(XLSX)