Impact of seed selection on subclonal reconstruction solutions

Project goal:

1. Quantify SRC result variability across SRC tools
2. Quantify SRC result variability across initializing random seed

Description

In this project, we have expanded the subclonal reconstruction pipeline https://github.com/uclahs-cds/pipeline-call-SRC to accept the output of multiple mutation callers.

We have integrated 4 additional mutation callers (MuTect2, SomaticSniper, Strelka2, Battenberg) by creating parsers that extract variant data from the different tools' output. The parsers can be found here https://github.com/uclahs-cds/tool-SRC-util.

Random Seed Selection

Generating random seeds:

Generate initial seed: head -c 4 /dev/urandom | od -An -tu4 => 3058353505

Generate 10 random seeds:

import random
random.seed(3058353505)
random.sample(range(0, 1000000), k=10)

Chosen random seeds:

Generated/chosen seeds: [51404, 366306, 423647, 838004, 50135, 628019, 97782, 253505, 659767, 13142]

Tumour Sample Data:

[HNSC] Head and Neck cohort: /hot/project/disease/HeadNeckTumor/HNSC-000084-LNMEvolution/data/

[CPCGENE] Prostate cohort: /hot/project/disease/ProstateTumor/PRAD-000005-293PT/

Tools:

sSNV-caller: mutect2, strelka2, somaticsniper

sCNA-caller: battenberg

src-tool: pyclone-vi, dpclust, phylowgs

Node selection:

• PyClone-VI: F32 (average run time 15s - 10min)
• DPClust: F32 (average run time 5min - 40min)
• PhyloWGS: F72 (average run time 7h - 24h)

sSNV-caller and sCNA-caller results:

sSNV-caller output:

/hot/project/disease/HeadNeckTumor/HNSC-000084-LNMEvolution/data/SNV/<sSNV-caller>/recsnv/vcfs/

sCNA-caller output:

/hot/project/disease/HeadNeckTumor/HNSC-000084-LNMEvolution/data/<sCNA-caller>/

Modes:

[HNSC]

• Single region mode (sr) (run on primary tumour only)
• Multi region mode (mr) (run on primary and metastatic tumours)

[CPGENE]

• Multi region mode (mr) (run on multiple regions of primary tumour)

Running The Pipeline

The pipeline is run for each sample and seed in both single region and multi region mode.

Templates: /hot/project/method/AlgorithmEvaluation/BNCH-000082-SRCRNDSeed/pipeline-call-src/templates/

Pipeline input files:

configs:

/hot/project/method/AlgorithmEvaluation/BNCH-000082-SRCRNDSeed/pipeline-call-src/run-<sSNV-caller>-<sCNA-caller>-<src-tool>/input/config/seed_<seed>.config

• 1 config per seed
• indicates src-tool choice and pipeline parameters
• indicates pipeline run output directory

yamls:

/hot/project/method/AlgorithmEvaluation/BNCH-000082-SRCRNDSeed/pipeline-call-src/<sSNV-caller>_<sCNA-caller>_yamls/

single-region mode (one primary tumor sample: sinlge-region/<sample_id>.yaml

multi-region mode (primary and metastatic tumor samples): multi-region/<patient_id>.yaml

• 1 yaml per patient
• path to sSNV-caller output
• path to sCNA-caller output

Submission script:

Log files:

/hot/project/method/AlgorithmEvaluation/BNCH-000082-SRCRNDSeed/pipeline-call-src/run-<sSNV-caller>-<sCNA-caller>-<src-tool>/logs/

Create conda environment with pyyaml and numpy installed. Execute submission script in logs directory and from activated conda environment.

submission script: <mode>_<sSNV-caller>_<sCNA-caller>_<src-tool>_submission_script.sh

Completed pipeline combinations:

• [x] Strelka2-Battenberg-PyClone-VI (sr/mr)
• [x] Strelka2-Battenberg-DPClust (sr)
• [x] Strelka2-Battenberg-PhyloWGS (sr)
• [x] SomaticSniper-Battenberg-PyClone-VI (sr/mr)
• [x] SomaticSniper-Battenberg-DPClust (sr)
• [x] SomaticSniper-Battenberg-PhyloWGS (sr)
• [x] Mutect2-Battenberg-PyClone-VI (sr/mr)
• [x] Mutect2-Battenberg-DPClust (sr)
• [x] Mutect2-Battenberg-PhyloWGS (sr)

Results:

Output files:

/hot/project/method/AlgorithmEvaluation/BNCH-000082-SRCRNDSeed/pipeline-call-src/run-<sSNV-caller>-<sCNA-caller>-<src-tool>/output/

License

Author: Anna Neiman-Golden(aneimangolden@mednet.ucla.edu), Philippa Steinberg(psteinberg@mednet.ucla.edu)

[This project] is licensed under the GNU General Public License version 2. See the file LICENSE.md for the terms of the GNU GPL license.

<one line to give the project/program's name and a brief idea of what it does.>

Copyright (C) 2021 University of California Los Angeles ("Boutros Lab") All rights reserved.

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.