checklist for a new study

[jjgao]

@ritikakundra do we have a checklist for a new study. Could you refine the list below?

• Does study name and study ID follow our convention? e.g. Tumor_Type (Institue, Journal Year); brca_mskcc_2015. For mixed studies, let's not use mixed and decide case by case.
• is study meta data complete? e.g. pmid, group of PUBLIC
• were all samples profiled with WES/WGS? If not, is gene panel file curated?
• with matched normal?
• when pushing data out from msk-impact, germline mutations should be removed (unless they are used in the paper)
• are oncotree codes of all samples curated; Cancer Type and Cancer Type Detailed needs to be added in addition to Oncotree Code
• clinical sample and patient data with meta files
• mutations data with meta files
• MAF is based on hg19
• MAF with 2 isoforms: uniprot and mskcc
• CNA data with meta files
• CNA segment data with meta files
• Expression data including z-scores with meta files
• Case-lists for all profiles.
• Passes Validation
• Manual checking (Niki or JJ): MSKCC or Triage Portal

[ritikakundra]

@jjgao I added meta-studies to all the data files and case lists. But I don't know if CNA, segment, and expression is a must for public studies, since most publications we see only have mutations data. The MAF is the minimum genomic data file required for a study to be created.

[jjgao]

@ritikakundra if data does not exist, you can also check it off. The purpose of a checklist is to check one by one. And we can also evolve this list over time.

To utilize this checklist, one option would be creating a pull request per study with this checklist in the PR.

[jjgao]

cc'ing @schultzn to see if he has anything to add.

[ritikakundra]

@jjgao ok. Added a few more things

[jjgao]

Thanks, @ritikakundra. Let's start to use this checklist when pushing out new studies?

[ritikakundra]

@yichaoS @rmadupuri Our checklist for any dataset that goes public.

[jjgao]

@ritikakundra @yichaoS @rnugraha let's also try to improve this checklist over time. For new studies, could you create a pull request per study and paste this checklist there? Thanks.

[schultzn]

looks good to me!

[rnugraha]

@sandertan you might also want to check this