SNP array analysis was performed on all 17 samples using Beadchip technology from Illumina, but with different chips depending on availability at the time of the analysis (Supplementary Table S3 ). In particular, two samples were analyzed using Illumina Human660W-Quad chip (655,246 SNPs), six samples using Illumina HumanOmniExpress v1 (730,525 SNPs), and nine samples using Illumina HumanOmni1S (1,185,076 SNPs). In all cases, raw data were processed with Illumina Genome Studio v2009 with the Genotyping module v1.1.9 to extract B-allele frequency (BAF) and log R ratio (LRR) values for each SNP.
SNP array data were analyzed using the R package ASCAT (Van Loo et al, 2010 (link)) to obtain loss-of-heterozygosity (LOH) and allele-specific copy number (CN) profiles from the BAF and LRR values. All samples were analyzed independently and treated as unpaired samples, using the germline genotype prediction functionality from ASCAT. In short, after loading BAF and LRR data, the germline genotype parameters were estimated and the data were segmented using the ASPCF algorithm. Next, ASCAT computed the most likely combination of CN states, total ploidy and percentage of aberrant cells. Circular genomic plots were created using Circos (Krzywinski et al, 2009 (link)).
SNP array data were analyzed using the R package ASCAT (Van Loo et al, 2010 (link)) to obtain loss-of-heterozygosity (LOH) and allele-specific copy number (CN) profiles from the BAF and LRR values. All samples were analyzed independently and treated as unpaired samples, using the germline genotype prediction functionality from ASCAT. In short, after loading BAF and LRR data, the germline genotype parameters were estimated and the data were segmented using the ASPCF algorithm. Next, ASCAT computed the most likely combination of CN states, total ploidy and percentage of aberrant cells. Circular genomic plots were created using Circos (Krzywinski et al, 2009 (link)).