Tumor-only-model

[RoniHaas]

Here is the benchmarking result of the newly added Mutect2 option --germline-resource:

Benchmarking

• Normal-paired-model :
  • sample: A-full-p2
  • vcf comparison result: /hot/pipeline/development/slurm/call-sSNV/unreleased/maotian-add-germline-resource/vcf_comparsion_Afull.txt

Option	Total SNPs	Unique SNPs and %	Shared SNPs and %
With germline resource	133,585	10; 0.0%	133,575; 100.0%
W. O. germline resource	133,667	92; 0.1%	133,575; 99.9%

- sample: CPCG0196
- vcf comparison result: /hot/pipeline/development/slurm/call-sSNV/unreleased/maotian-add-germline-resource/CPCG_vcf_comparsion_paired.txt

Option	Total SNPs	Unique SNPs and %	Shared SNPs and %
With germline resource	34,747	30; 0.1%	34717; 99%
W. O. germline resource	43,276	8559; 19.8%	34717; 80.2%

• Tumor-only-model :
  • sample: anaplastic thyroid cancer sample - MDA10
  • vcf comparison result: /hot/pipeline/development/slurm/call-sSNV/unreleased/maotian-add-germline-resource/vcf_comparsion.txt

Option	Total SNPs	%Unique SNPs	%Shared SNPs
With germline resource	122,005	1488; 1.2%	120517; 98.8%
W. O. germline resource	3,552,165	3431648; 96.6%	120517; 3.4%

 - sample: CPCG0196
- vcf comparison result: /hot/pipeline/development/slurm/call-sSNV/unreleased/maotian-add-germline-resource/CPCG_vcf_comparsion_tumor_only.txt

Option	Total SNPs	Unique SNPs and %	Shared SNPs and %
With germline resource	149,053	2737; 1.8%	146316; 98.2%
W. O. germline resource	189,592	3165672; 95.6%	146316; 4.4%

• Multiple-tumor-model :
  • sample:anaplastic thyroid cancer sample - ATCWGS21
  • vcf comparison result: /hot/pipeline/development/slurm/call-sSNV/unreleased/maotian-add-germline-resource/vcf_comparsion_ATCWGS.txt

Option	Total SNPs	%Unique SNPs	%Shared SNPs
With germline resource	13,104	2; 0.0%	13,102; 100%
W. O. germline resource	13,561	459; 3.4%	13,102; 96.6%

• Comparison between tumor-only and paired Mutect2
  • sample: CPCG0196
  • vcf comparison result: /hot/pipeline/development/slurm/call-sSNV/unreleased/maotian-add-germline-resource/CPCG_vcf_compare_paired_tumor_only.txt

Option	Total SNPs	%Unique SNPs	%Shared SNPs
Paired Normal-Tumor	34,747	14138;40.7%	20,609; 59.3%
Tumor-only	149,053	128,444; 86.2%	20,609; 13.8%

Originally posted by @maotian06 in #80

[RoniHaas]

Originally posted by @maotian06 in #80

Hey @maotian06,
I am interested in running the call-sSNV in a Tumor-only-model but not so sure how to do that. Can you please help with some details?

[tyamaguchi-ucla]

@RoniHaas this PR might help - https://github.com/uclahs-cds/pipeline-call-sSNV/tree/maotian-update-README

[RoniHaas]

@RoniHaas Any details of your samples? I will close this issue and move it to the discussion.

@RoniHaas this PR might help - https://github.com/uclahs-cds/pipeline-call-sSNV/tree/maotian-update-README

Thank you! All I need to do is to use this yaml template https://github.com/uclahs-cds/pipeline-call-sSNV/blob/maotian-update-README/input/example-test-tumor-only.yaml
@maotian06 @tyamaguchi-ucla
Should I change the config file?

[maotian06]

You can use the newest release here: /hot/software/pipeline/pipeline-call-sSNV/Nextflow/release/4.0.0-rc.1/
For input.YAML, you can copy config/template.config and this YAML to your own directory and change it: /hot/software/pipeline/pipeline-call-sSNV/Nextflow/release/4.0.0-rc.1/input/example-test-tumor-only.yaml

[maotian06]

@RoniHaas
Any details of your samples?
I will close this issue and move it to the discussion.

[RoniHaas]

.I would run the latest call-gSNP multi-sample mode to realign the 4 samples all together if you haven't.

By 4 samples, do you mean the 3 radiosensitive and one radioresistant?

If I recall, they are biological replicates, right (i.e. same experiment but cultured separately in the same environment -> different mutation distributions)?

The definition of biological or technical is a bit tricky here I think. In terms of mutations, I tend to consider them as technical replicates. They were taken from the same plate and then were grown separately before extraction.

[tyamaguchi-ucla]

.I would run the latest call-gSNP multi-sample mode to realign the 4 samples all together if you haven't.

By 4 samples, do you mean the 3 radiosensitive and one radioresistant?

Yes. (I'm not sure if 2TB /scratch space is enough though..)

If I recall, they are biological replicates, right (i.e. same experiment but cultured separately in the same environment -> different mutation distributions)?

The definition of biological or technical is a bit tricky here I think. In terms of mutations, I tend to consider them as technical replicates. They were taken from the same plate and then were grown separately before extraction.

Yeah, so they were grown separately so you would expect different sets of somatic mutations from the 3 samples. The definition in this context is here Q1 - https://confluence.mednet.ucla.edu/display/BOUTROSLAB/Dataset+registration+-+FAQ

@maotian06 is it possible to enable the tumor-only mode and multi-sample mode together in call-sSNV? It should be.

[maotian06]

@RoniHaas One concern is that you will get many false-positive SNVs using the tumor-only model.
So I wondered if you are using the cell line data, is it possible to generate a panel of normal samples to a normal.bam and then run it...

[maotian06]

@tyamaguchi-ucla I tested with Ardy's sample with 7 bams, it works, so I think the multi-sample would be enough for the Roni's case.

Yes! Multi-sample and Tumor-only can work together! @RoniHaas

[RoniHaas]

@RoniHaas One concern is that you will get many false-positive SNVs using the tumor-only model. So I wondered if you are using the cell line data, is it possible to generate a panel of normal samples to a normal.bam and then run it...

Since these are cell lines, I am not sure that creating a panel of N samples from some patient samples that we have in the lab is the best solution. I also only need an evaluation of the number of sSNVs in these samples. I am not going to use the positions probably. Maybe in my case single-tumor mode is good enough?