Application talk
Abstract
Summary: Through collaboration on open source software, pharma, commercial software development and cancer research institutions have proven successful in enhancing the cBioPortal platform by optimizing and extending it with new features. Approximately one year ago the popular cBioPortal for Cancer Genomics was made open source. In this last year its development community has grown and the platform has been extended with many new features. Here we detail some of the contributions The Hyve (Utrecht) has made to the platform, in collaboration with Dana Farber Cancer Institute (Boston), Memorial Sloan Kettering Cancer Center (New York) and Boehringer Ingelheim (BI RCV). The contributions can roughly be divided into three categories: (1) improvement of the data loading pipeline, (2) new data analysis features, and (3) optimizations of the front end. In the data loading pipeline we have introduced a strict separation between the validation step and the loading step. This “separation of concerns” design principle makes the code easier to understand and maintain and simplifies the process of adding new datasets to a local cBioPortal installation. Special effort was spent on making the validator easy to use, which is exemplified by clearer error messages and the generation of a HTML validation report. In the front end we added a whole new pan-cancer view for studies comprising multiple cancer types, added new query options in the Study overview page and added new visualizations to the query results page to support better enrichment analysis of expression (mRNA, Proteins) and co-occurrence (copy number, mutations). We have also implemented integration documentation from the Wiki or Git, and made the portal more customizable (logo, headers, news and FAQ), which is very important for open source software. Last but not least, we have optimized the loading times of the portal to be able to host larger studies, focusing on the most used pages in the application. In the query results page we have successfully shortened the loading times of various analyses.
Authors
Pieter Lukasse, The Hyve, Netherlands
Sjoerd van Hagen, The Hyve, Netherlands
Sander de Ridder, The Hyve, Netherlands
Fedde Schaeffer, The Hyve, Netherlands
James Lindsay, Dana Farber Cancer Institute, United States
Jianjiong Gao, Memorial Sloan Kettering Cancer Center, United States
Benjamin Gross, Memorial Sloan Kettering Cancer Center, United States
Zachary Heins, Memorial Sloan Kettering Cancer Center, United States
Priti Kumari, Dana Farber Cancer Institute, United States
Adam Abeshouse, Memorial Sloan Kettering Cancer Center, United States
Hongxin Zhang, Memorial Sloan Kettering Cancer Center, United States
Robert Sheridan, Memorial Sloan Kettering Cancer Center, United States
Onur Sumer, Memorial Sloan Kettering Cancer Center, United States
Stuart Watt, Princess Margaret Cancer Centre, United States
Chris Sander, Dana Farber Cancer Institute, United States
Nikolaus Schultz, Memorial Sloan Kettering Cancer Center, United States
Ethan Cerami, Dana Farber Cancer Institute, United States
Yichao Sun, Memorial Sloan Kettering Cancer Center, United States