Incorporate Aims feedback and redo the 1 page aims

[cvraut]

so I finished receiving all the feedback from the reviewers:

Mattew O'Meara

Aims:

• Format / Citations: ok
• Technical details: Ok
• Framing of the specific problem statement:
  • Clustering phenotypes is reasonable but you need more justification for why there should be 10 clusters. E.g., If you sort individuals into 5 choose 3 bins, where would you put someone with high scores in all 5 quantitative traits? In general for quantitative traits you can have an unbounded number of clusters, but in practice it's limited by noise in the data to reliably distinguish clusters.
  • Since you are focusing on exonic variants, consider mapping them to the actual protein structure predict functional impact. For example, measure the TM-align between AlphaFold2 structures of the reference and variants.
• Clear description of the data: Needs clearer statement of the scope of the data (e.g. how many samples/individuals/pedigrees, how many variants/genes)
• Clear description of the deliverables of the aims:
  • Not clear how this work would complement or differentiate from other GWAS for rare variants for MetS

Lana Garmire

• SKAT-O test is mentioned in the introduction but not used in the aims. Not sure why it needs to be mentioned then since it looks irrelevant.
• The 5 traits need to be cited and listed what they are.
• While I quite like the “leave-one-out” trial to evaluate the effect of a variant, the caveat here is that this variant does not interact with any other variant, and this not the best assumptions. Need to do prior testing to assure that this assumption holds.

Evan Snitkin

• Have prior studies for MetS found most GWAS hits in coding regions versus non-coding? If not, then it might be important to justify your decision to focus on coding.
• I am not familiar with LOVO studies in this context, but I wonder if burden tests based on functional domains/motifs, along the lines of Matt’s comment might be worth considering as well
• Not necessary for the abstract, but for your proposal it will be important to consider your power to detect rare variants for this type of complex phenotype, and provide conceptual support from similar studies.

---

PLAN of attack:

• Step 1: would be to try to address each one on my own
  • note down a couple bigger issue things.
• Step 2: consolidate with Antonio on the big picture items

[cvraut]

finished v2 a couple of days ago. need to think about v3 after finishing report

[cvraut]

Aims page got refactored a bit, but I'm done now