Fig. 1
From: A multivariable approach for risk markers from pooled molecular data with only partial overlap
Scenarios appearing in the analysis of consortial data based on two studies after imputation. a. Illustration of SNP data for all individuals in a study. Every row contains all SNP data for one individual and each column represents the data for one SNP and all individuals. b. A perfect world: Both studies cover the same SNPs for all individuals (full). c. Reality: Differential coverage of SNPs in both studies. All SNPs in study B are a real subset of the SNPs in study A. An ideal analysis can use all applicable information (indicated by red for reduced). d. Reality: Differential coverage of SNPs in both studies as in Fig. 1c. In a complete case analysis, all information from study B is dropped (indicated by part for partial)