370k array

The following genotyping arrays were used for genotyping the study cohorts: Illumina 370K array for NFBC1966, Illumina 610K array for COROGENE and GenMets, custom generated Illumina 670K array for YFS and HBCS, Illumina 770K array for PredictCVD, Illumina HumanCoreExome-12v1-0 for FR and Affymetrix GeneChip Human Mapping 500K Array set for DGI. The quality control procedures included removing closely related individuals (PI_HAT >0.1) by analyzing pairwise IBD relationships for all individuals in five Finnish discovery cohorts together and have been described previously in detail [29] (link), [32] (link). The imputation of non-pseudoautosomal chrX variants into the study cohorts was performed on the cleaned data in each cohort using IMPUTE version 2.2.2 [33] (link), [34] (link). The reference panel used in the imputation was the integrated variant set release (v3) released in March 2012 (http://mathgen.stats.ox.ac.uk/impute/data_download_1000G_phase1_integrated.html). The data were split into genomic regions of ∼5 Mb (with 250 kb (DGI) or 1 Mb (other cohorts) buffer region), using effective population size of 20000 (DGI) or 11418 and k value of 80.

We downloaded data from GEO (accession GSE32649) from a study of hepatocellular carcinoma [10] (link), in which both cancer and surrounding normal tissues from 86 patients were analyzed on an Illumina 370K array. We ran J-LOH on the normal samples only, with a fixed TPM assuming that aberrant events had a mean length of 600 markers (about 20 Mb) and covered 0.1% of the genome. We visually inspected posterior probability plots and array data of matched tumor samples for aberrant regions with matching boundaries and event type.

The Cilento isolate comprise three villages from the South of Italy; Campora, Cardile, and Gioi. Pedigree, phenotypic, and genetic data have previously been gathered as part of the Cilento Study. A pedigree structures which connects all three village has been reconstructed from parish records. The three villages have been shown to represent characteristics of population isolates intermediate between the large isolate population of Iceland^{62 (link)} and the highly isolated Hutterite population^{63 (link),64 (link)}. Aggregating over the three villages, we have a pedigree of 7,585 members including 1,444 genotyped members. The high quality of the reconstructed genealogy in Cilento makes it an appropriate tool for simulating a realistic example of data from an isolated population. Individuals from Campora and Cardile have been genotyped on an Illumina 370 K array, whilst individuals from Gioi have been genotyped on an Illumina HumanOmniExpress array. Deep phenotyping has been performed in Cilento for a range of anthropometric, cardiometabolic, and haematological traits. For the purposes of this study, we have concentrated on phenotypes that have been often analyzed in the literature of both other population isolates and in samples of unrelated individuals (Supplementary Table 1).