Proceedings of the Critical Assessment of Massive Data Analysis conferences: CAMDA 2011 (Vienna, Austria) and CAMDA 2012 (Long Beach, CA USA)

Abstract

CAMDA has now evolved from its origins at Duke University in the year 2000, founded by Simon Lin and Kimberly Johnson, to an international conference of renown that has been affiliated with ISMB/ECCB since the 2008 meeting in Vienna, and which is now a regular official Satellite Meeting of the ISMB Conference. Since 2011, proceedings are published Open Access in partnership with Systems Biomedicine. At the CAMDA conferences, alternative analyses of annually set Contest Datasets are discussed which have been submitted by different research teams. Selected contributions are collected in the special proceedings volume presented here, including the analyses by the teams of Sol Efroni and Djork-Arné Clevert, which were chosen as the best contributions by secret vote of the delegates of CAMDA 2011 and 2012, respectively. By design, CAMDA analysis goals and the competition are very openended, which is a distinguishing feature of the contest. CAMDA can therefore take on the most challenging data sets. Over the last few years, both complex multi-track data sets and unusually large measurement series have been featured. For CAMDA 2011, the Glioblastoma multiforme subset of The Cancer Genome Atlas (TCGA) had been identified as a particularly interesting challenge. It is unusual in that it provides publicly, for several hundred patients, profiles of gene transcript expression (435 cancer patients vs 11 controls), miRNA expression (426 tumor samples vs 10 controls), genomic DNA methylation (256 tumor samples vs a control), and copy number variation (465 tumor samples vs 430 controls, including 402 matched normals), which are complemented by a variety of clinical parameters and survival outcomes. Sometimes, additional results are available from alternative technologies/platforms. The data can be downloaded at different abstraction levels, from raw (Level 1, publicly available for some platforms) via normalized (Level 2) to processed (Level 3), also facilitating integration and participation by non-domain experts. Typical questions of interest include the investigation to what degree the integration of such large heterogeneous repositories can improve our understanding of complex biomedical questions. For the 2012 contest, a subset of the Japanese Toxicogenomics Project focusing on liver proved the most popular contest data set, containing over 21,000 arrays for rats treated with mainly human drugs and profiled using the Affymetrix RAE230_2.0 GeneChip. Expression profiles (raw and processed data) were complemented by drug information and pathology data that had been compiled by Weida Tong of the US FDA. Typical questions of interest include whether a better prediction of liver-toxicity from animal experiments can be achieved, and to what degree animal studies could be replaced by in-vitro assays. In fact, the data set sparked such interesting discussions at the conference, that it has been offered again also in the following years. As in past years, the CAMDA conferences have been graced by fantastic keynotes complementing the competition program, including presentations by Terry Speed (WEHI, Australia), John Storey (Princeton, USA), and Stefano Volinia (Ferrera, Italy) at the meeting 2011 in Vienna, while in 2012 keynote lectures by Olga Troyanskaya (Princeton) and Weida Tong (FDA) provided further stimulating insight and discussions. Some of these talks are available on the respective pages of the conference website www.camda.info. Looking forward, the CAMDA 2013 proceedings will be published shortly in this series of Special Volumes, and the Contest is open for the 2014 conference to be held at Boston, USA. On behalf of the conference co-chairs, we thank all the contributors to making the meetings such an intellectually stimulating lively exchange, which has been the foundation for the long running success of the conference. Last but not least, we thank Landes Bioscience for their long-term support of the conference.