Cardiogram+C4D sex stratified analysis plan for X-chr and autosomes (Jan 2024)

Rational

The impressive progress in understanding the genetic architecture of CAD that has been achieved in recent months by collaborative efforts from CARDIoGRAMplusC4D and the Million Veterans Project, enhanced by individual studies as well as international biobanks (e.g. Japan Biobank, UK Biobank), has resulted in unprecedented data availability in the field. Genome-wide summary data (i.e. association statistics) from these efforts is currently generated for all autosomal chromosomes. Combining data from these resources in a transethnic meta-analysis (and related analyses), and ex-panding upon the contributing cohorts, in advance of public data release is a timely priority that would further enhance our understanding of the transethnic genetic architecture of CAD

Primary goal

• To undertake the X-chromosome ancestry specific and multi-ancestry meta-analysis using all cohorts currently part of the Card+C4D extension.
• To undertake the sex stratified autosomal-chromosome ancestry specific and multi-ancestry meta-analysis using all cohorts currently part of the Card+C4D extension.

Case/Control definition

Note: We anticipate using the CAD definition that was used for the autosomal chromosome analysis for cohort that have already provided sex combined analysis (filled out the specific definition of CAD that was used, this will help standardize defi-nition of CAD for all cohorts). Below are detailed, case/control adjudication for new study undertaking the analysis.

Case definition

Coronary artery phenotypes (preference is for the following or minor variation to it – please do contact us to discuss any concerns or variations). We wish to avoid overly soft phenotypes as pre-vious work has established the limited value of such cases. Consequently, for de novo studies un-dertaking new analyses, cases should adhere to the following as closely as possible. However, in-dividual studies should use insights into their data resource to adapt appropriately (and provide de-tails of the phenotype provided).

• Confirmed MI (see below)
• Over 50% stenosis in at least one coronary vessel at angiography with validation from hospital records
• Validated history of percutaneous transluminal coronary angioplasty (PTCA) or coronary artery bypass graft surgery (CABG)
• Validated angina, defined as symptoms + confirmation from at least one non-invasive provocation test e.g. scintigraphy or exercise treadmill test (angina cases based only on self-report angina alone should not be included).
• Death confirmed to be due to CAD or highly likely to be due to CAD
• Hospitalization or death recorded under ICD10 I21-I25

Control definition

• All participants not defined as coronary cases using the most liberal ‘Soft’ definition to be “non-cases” in the analysis. Note: Please specify if you are using population-based controls

Association testing

For each sex and all chromosomes, including the X-chr, please test additive models using logistic regression, accounting for genotype imputation uncertain-ty (i.e. SNP probability or dosage). For studies that have already conducted analyses, please dis-cuss the models used (for example, some studies may have already adjusted for age and this is not considered a substantial deviation from the analysis plan). Use study appropriate software to ac-count for (or exclude as appropriate) relatedness. Detail analysis plan can be found at:

• Sex stratified autosomal chromosome: Sex_stratified_autosomal_chr.md
• Sex Stratified X-chromosome: Xchr_analysis_plan.md

Computational tools

PLINK (https://www.cog-genomics.org/plink/2.0/)

REGENIE (https://rgcgithub.github.io/regenie/)

SAIGE (https://saigegit.github.io/SAIGE-doc/)

XWAS (https://github.com/KeinanLab/xwas-3.0)

Contact:

If you have any question or concern please contact any of the following: Catherine Tcheanjieu: Catherine.tcheandjieu@gladstone.ucsf.edu, Silke Szymczak: silke.szymczak@uni-luebeck.de; Stravoula Kanoni s.kanoni@qmul.ac.uk