Increased availability of drug response and genomics data for many tumor cell lines has accelerated the development of pan-cancer prediction models of drug response. However, it is unclear how much between-tissue differences in drug response and molecular characteristics may contribute to pan-cancer predictions. Also unknown is whether the performance of pan-cancer models could vary by cancer type. Here, we built a series of pan-cancer models using two datasets containing 346 and 504 cell lines, each with MEK inhibitor (MEKi) response and mRNA expression, point mutation, and copy number variation data, and found that, while the tissue-level drug responses are accurately predicted (between-tissue ρ = 0.88-0.98), only 5 of 10 cancer types showed successful within-tissue prediction performance (within-tissue ρ = 0.11-0.64). Between-tissue differences make substantial contributions to the performance of pan-cancer MEKi response predictions, as exclusion of between-tissue signals leads to a decrease in Spearman's ρ from a range of 0.43-0.62 to 0.30-0.51. In practice, joint analysis of multiple cancer types usually has a larger sample size, hence greater power, than for one cancer type; and we observe that higher accuracy of pan-cancer prediction of MEKi response is almost entirely due to the sample size advantage. Success of pan-cancer prediction reveals how drug response in different cancers may invoke shared regulatory mechanisms despite tissue-specific routes of oncogenesis, yet predictions in different cancer types require flexible incorporation of between-cancer and within-cancer signals. As most datasets in genome sciences contain multiple levels of heterogeneity, careful parsing of group characteristics and within-group, individual variation is essential when making robust inference.

MOTIVATION: DNA copy number variants (CNVs) are gains and losses of segments of chromosomes, and comprise an important class of genetic variation. Recently, various microarray hybridization-based techniques have been developed for high-throughput measurement of DNA copy number. In many studies, multiple technical platforms or different versions of the same platform were used to interrogate the same samples; and it became necessary to pool information across these multiple sources to derive a consensus molecular profile for each sample. An integrated analysis is expected to maximize resolution and accuracy, yet currently there is no well-formulated statistical method to address the between-platform differences in probe coverage, assay methods, sensitivity and analytical complexity. RESULTS: The conventional approach is to apply one of the CNV detection ('segmentation') algorithms to search for DNA segments of altered signal intensity. The results from multiple platforms are combined after segmentation. Here we propose a new method, Multi-Platform Circular Binary Segmentation (MPCBS), which pools statistical evidence across platforms during segmentation, and does not require pre-standardization of different data sources. It involves a weighted sum of t-statistics, which arises naturally from the generalized log-likelihood ratio of a multi-platform model. We show by comparing the integrated analysis of Affymetrix and Illumina SNP array data with Agilent and fosmid clone end-sequencing results on eight HapMap samples that MPCBS achieves improved spatial resolution, detection power and provides a natural consensus across platforms. We also apply the new method to analyze multi-platform data for tumor samples. AVAILABILITY: The R package for MPCBS is registered on R-Forge (http://r-forge.r-project.org/) under project name MPCBS. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.