CNApp: a web-based tool for integrative analysis of genomic copy number alterations in cancer

Copy number alterations (CNAs) are a hallmark of cancer. Large-scale cancer genomic studies have already established the CNA landscape of most human tumor types and some CNAs are recognized as cancer-driver events. However, their precise role in tumorigenesis as well as their clinical and therapeutic relevance remain undefined, thus computational and statistical approaches are required for the biological interpretation of these data. Here, we describe CNApp, a user-friendly web tool that offers sample- and cohort-level association analyses, allowing a comprehensive and integrative exploration of CNAs with clinical and molecular variables. CNApp generates genome-wide profiles, calculates CNA levels by computing broad, focal and global CNA scores, and uses machine learning-based predictions to classify samples by using segmented data from either microarrays or next-generation sequencing. In the present study, using copy number data of well-annotated 10,635 genomes from The Cancer Genome Atlas spanning 33 cancer subtypes, we showed that patterns of CNAs classified tumor subtypes according to their tissue-of-origin and that broad and focal CNA scores correlated positively in those samples with low levels of chromosome and arm-level events. Moreover, CNApp allowed the description of recurrent CNAs in hepatocellular carcinoma further confirming previous results identified using other methods. Finally, we established machine learning-based models to predict colon cancer molecular subtypes and microsatellite instability based on broad and focal CNA scores and specific genomic imbalances. In summary, CNApp facilitates data-driven research and provides a unique framework to comprehensively assess CNAs and perform integrative analyses that enable the identification of relevant functional implications. CNApp is hosted at http://bioinfo.ciberehd.org/CNApp.