Greetings folks. We haven't done as good of job as we should have here keeping you up to date.

Batches 200-209 will be 9 hours long
Batches 210 and up will be 10 hours long

For each gene, we must have at least 100,000 structures and an average of approx 120,000. We start each gene by sending out 4 'seed' workunits for the gene. When 10 results for the workunit has come back, validation will be attempted. If successful, then we calculate the estimated number of structures that will be produced for the workunit. We then generate additional 'child' workunits for the gene so that enough structures will be generated. Each 'seed' is responsible for producing enough 'children' to produce 30,000 structures. Thus the four seeds produce the required 120,000 structures by creating additional workunits.

Example of how this works:

Workunit: R00221_00cd37b956c7d11e0230718f3dc4e1d7_00

(FYI: R00221 -> means that this is from batch 221)

19 copies are sent out. After the first 10 are received and validated we determine that 3,138 structures have been produced with these 10 results. We estimate that the workunit (when all results are returned) will produce 5,366 structures ( (3,138/10)*90%*19 ). Thus we estimated that we need 30,000/5,366 = 5.6 workunits (or 6) so an additional 5 workunits will be created. These workunits will be named:

R00221_00cd37b956c7d11e0230718f3dc4e1d7_00_000
R00221_00cd37b956c7d11e0230718f3dc4e1d7_00_001
R00221_00cd37b956c7d11e0230718f3dc4e1d7_00_002
R00221_00cd37b956c7d11e0230718f3dc4e1d7_00_003
R00221_00cd37b956c7d11e0230718f3dc4e1d7_00_004

Changing the time limit for the workunits only changes the number of structures produced in the 'seed' workunit. This will then change the number of additional workunits generated.