Additional Benchmarking to be added to README?

[sorelfitzgibbon]

/hot/user/sfitzgibbon/PCAWG-63/benchmark-call-ssnv

Input was 30 tumor/normal pairs from the PCAWG-63 dataset, chosen because they had complete pipeline-call-sSNV results. Default v8.0.0 F72.config resources were used.

Sample IDs:

• /hot/user/sfitzgibbon/PCAWG-63/benchmark-call-ssnv/benchmark-30-tumor.txt
• /hot/user/sfitzgibbon/PCAWG-63/benchmark-call-ssnv/benchmark-30-normal.txt

Nextflow trace files:

• /hot/user/sfitzgibbon/PCAWG-63/benchmark-call-ssnv/PCAWG-trace-files

Per sample mean coverage:

• /hot/user/sfitzgibbon/PCAWG-63/benchmark-call-ssnv/benchmark-30-tumor-meanCoverage.txt
• /hot/user/sfitzgibbon/PCAWG-63/call-ssnv-benchmark/benchmark-30-normal-meanCoverage.txt

Sample coverage

type	min	max	average
tumor	24x	71x	44x
normal	24x	50x	32x

Runtimes per sample (summed over parallel jobs)

Tool	min	max	average
MuSe	1h 54m	4h 40m	3h 6m
Mutect2			68h 10m
SomaticSniper			13h 15m
Strelka2			5h 30m

Cpus per sample

Tool	min	max	average
MuSe	1174%	1191%	1185%
Mutect2			4733%
SomaticSniper			291%
Strelka2			1121%

CPUhours

Tool	average
MuSe
Mutect2
SomaticSniper
Strelka2

Memory (max rss)

Tool	min	max	average
MuSe	33.3 GB	46.8 GB	39.1 GB
Mutect2			27.7 GB
SomaticSniper			4.41 GB
Strelka2			3.32 GB

[sorelfitzgibbon]

Had mistakes, will repost later