CAMDA Challenge Takes on CFS

Research teams convened atDuke University on June 8-9, 2006, to share their analyses of data from 227 CFS patients and control subjects identified in the Wichita population survey. The two-day conference was the culmination of the 2006 Critical Assessment of Microarray Data Analysis (CAMDA), an annual event that brings together multidisciplinary teams of scientists who use cutting-edge computational techniques to understand large, complex datasets. This year’s challenge invited teams to present their analyses of data compiled by the Centers for Disease Control and Prevention’s (CDC’s) Chronic Fatigue Syndrome (CFS) Research Group. It was the largest dataset ever made available to CAMDA teams and was unique in that it combined clinical data with information about gene expression, genetics and proteins.

Eleven groups representing five countries and 10 institutions presented their findings at the meeting. The papers selected for oral presentation and poster session abstracts can be found at http://www.camda.duke.edu/camda06/papers/.

Several of the presentations replicated the approaches and findings reported by four other teams that examined the dataset in another challenge, the CFS Computational Challenge (C3) sponsored by CDC.*

Two other CAMDA groups attempted to tease out differences between CFS and depression, showing that CFS and major depressive disorder fit different biologic models ( Seoul National University ) and that different single nucleotide polymorphisms (SNPs) are associated with CFS compared to depression (Oak Ridge National Laboratory).

A group from the University of California at Los Angeles used a novel weighted network analysis and proposed the gene Forkhead Box NI (NM_003593) as a candidate in CFS. The UCLA group won the Best Presentation award, based on a vote of meeting participants.

In all, the differences in proteins and gene expression produce some interesting candidates in looking for a CFS biomarker, but most groups reported that these data are more meaningful when integrated with the clinical data. With more extensive knowledge of the polymorphisms associated with CFS, genetic information will be useful in helping to define an individual’s risk of developing the illness and in confirming the clinical CFS diagnosis.

Groups at CAMDA and involved in C3 showed that data from CFS patients and control groups from the Wichita study can be stratified into numerous subgroups, depending on the factors selected by individual research teams. The different groupings of subjects has contributed to identification of a set of candidate genes that warrant further study as possible biomarkers and clues to the pathophysiology of CFS. CDC’s research group is collaborating with several of the CAMDA participants in hopes of validating findings. These computational/statistical/mathematical approaches to the data offer very promising advances in our understanding of CFS.

The CFIDS Association of America was pleased to sponsor both CAMDA and the reporting meeting for C3 held in September 2005 at Cold Spring Harbor Laboratories.

*The results of C3 were published in the April issue of Pharmacogenomics in a series of 14 articles. The publications warranted a press briefing at CDC, available at http://www.cdc.gov/od/oc/media/transcripts/t060420.htm , generated considerable media coverage—and sparked debate—by national and local news outlets and high-profile research publications including Science, Nature and the Journal of the American Medical Association. Watch the CFIDSLink and the Chronicle for more on this research and coverage of it in the press.