Post date: Jan 25, 2016 4:57:31 PM
Present on call Lies, Yohan, Brian, Deb, Soumitra, Louise, David Strachan, Steve Rennard, Kim, Marike
Discussion about the ICGC case-control meta-analysis paper.
- Replication of results: Kim re-performed analysis (independent analysis of top results using the regression results from each cohort), slight differences in results found - Brian has been investigating, and found that results from one cohort were truncated in the file sent to Kim, and other results were QC'd out by EasyQC -- these explain the differences, and the results can be replicated from the clean files.
- Target journal: might be reasonable to send an inquiry to Nature Genetics, one downside would be a delay if it does get sent out for review and then rejected. Some good suggestions from the group to make it more interesting -- change the title to highlight overlap with lung function and novel loci; non-overlap with asthma; lack of more support for Alpha-1 (and why). If not there, then AJRCCM reasonable (Nat Communications similar impact factor, an alternative).
- Sharing of results: required of Nat Genet, also would be relevant for future studies. Many groups are familiar and accepting of this (e.g. all CHARGE, also BiLEVE), also most cohorts (CODPGene, ECLIPSE, SPIROMICS) will be publicly deposited, perhaps making this point moot. Many different potential mechanisms to share; would assume only summary statistics (not cohort specific results) of 1) Beta / SE, or just P-values, 2) available to anyone (no restrictions), requiring some form of registration or access (dbGap, UK Biobank), or restrictive - requiring ancillary study approval of all cohorts. Generally some sharing reasonable to most people; will do some more information gathering about this process / approval (e.g. CHARGE?) and also get add'l input from investigators.
- For future calls; next 'phase' of case control associations, lung function and severe analyses, and emphysema analysis.