There was a news release a couple of days ago accompanying a paper in PNAS concerning a possible large effect of time of day on the risk of initial viral infection. Google "news reddy bmal1 influenza virus" for more info. This of course has come too late to have any effect on this DREAM challenge, but it certainly needs to be taken into account when any new scientific research on this topic is contemplated. For one thing the time of day of each observation has not been included in the public Duke/VA data already released. It would also seem likely that most or all of the observations were taken during the middle of the day or at least a fixed time, so the full effect of the circadian clock on susceptibility is not going to be seen. Note that the gene symbol "Bmal1" used by the Reddy group at Cambridge is no longer the approved one, which is "ARNT". This line of research is still very much in the early stages, but there is already a substantial literature on time of day and seasonal effects on viral infection in other animals. See the book "Ecoimmunology" (Oxford 2012) edited by Gregory Demas at Indiana U and Randy Nelson at Ohio State for more info. I am continuing to work on my own approach to building prediction models for gene expression based on microarray data, and should have some actual results in the not too distant future, though they would not be directly comparable to the ones submitted for this competition. Also though while my methods are mostly quite simple mathematically, it would not be feasible to train other researchers in using them in the short time frame of this type of competition.