Expected YAML input for multiple samples

[maotian06]

Two things want to achieve here:

1. include the multiple samples in the call-sSNV - mutect2 algorithms
2. update the input to YAML.

Here is the expected YAML file:

---
# Sample inputs
sample_name: ''

# algorithm and reference used
algorithm : [] // 'somaticsniper', 'strelka2', 'mutect2'
reference : ''

# options for specical tumor mode
tumor_only_mode: false
multi_sample_mode: true

# path to BAMs
normal:
  -[id1, path/to/1.bam]
  -[id2, path/to/2.bam]
tumor:
  -[id1, path/to/1.bam]
  -[id2, path/to/2.bam]

@tyamaguchi-ucla @yashpatel6 @zhuchcn

[yashpatel6]

I would suggest keeping the algorithm and reference params in the config file and keep only the sample-specific params in the YAML. For tumor_only_mode and multi_sample_mode, if these will be consistent across most samples, you can have a default value in the config file and add a comment in the YAML that if specified in a sample YAML, they'll over-write the config file param for the specific sample. So maybe something like:

---
# Sample inputs
sample_name: ''

# options for specical tumor mode
# If included in sample YAML, the defaults in the config file will be overwritten
tumor_only_mode: false
multi_sample_mode: true

# path to BAMs
normal:
  -[id1, path/to/1.bam]
  -[id2, path/to/2.bam]
tumor:
  -[id1, path/to/1.bam]
  -[id2, path/to/2.bam]

[maotian06]

Okay! I will add the algorithm and reference in the YAML! But I also want to keep the algorithm and reference in the template.config, so the users don't need to define them when they are the same across the cohort.
In this case, their YAML files will be really simple. Something like:

---
sample_name: ''
normal:
  -[id1, path/to/1.bam]
  -[id2, path/to/2.bam]
tumor:
  -[id1, path/to/1.bam]
  -[id2, path/to/2.bam]

[tyamaguchi-ucla]

Hi guys,

Some general thoughts. Do we want to use list in pipeline YAML inputs? If so, we might have to deal with magic numbers so I think dictionary (key:value) would be easier to use and more flexible (even with nesting)

We used to use these YAML formats - https://confluence.mednet.ucla.edu/pages/viewpage.action?pageId=77029703 (legacy call-gSNP) and https://confluence.mednet.ucla.edu/display/BOUTROS/Alignment+Pipeline+How-To (legacy align-DNA)

[yashpatel6]

Right, I think we do want to use dictionaries. Mao and I talked recently and decided to switch it to a dictionary format, similar to the example in call-FusionTranscript.

[maotian06]

@tyamaguchi-ucla I cannot open the link you provide...
YES! Yash showed me how to use dictionary in YAML files and nextflow, and I am trying it out now. Will update the result here!

[tyamaguchi-ucla]

@maotian06 you can ask Julie R for access.