1) What is the difference between rounds. Will there be new data in each round? 2) In each round a number of submissions are made. I can, as the last submission, take the best performing clustering from each of my previous submissions (across the different networks) to optimise my NS score. Should I do that, or does the challenge already take this into account? Just to be sure, I am allowed to use different parameters/methods for different networks, within a single submission? 3) Is NS the main determinant of the leaderboard? 4) All teams are exploring granularity space by themselves. Has it been considered to make some sort of information available to participants, e.g. a size distribution of clusters deemed significant across all participants, possibly separated across networks? The challenge is really well-organised, and the organisers have been very helpful and quick to answer questions. Thanks!

Created by Stijn van Dongen speedwell
Hi Stijn Thanks for the feedback, you are raising some very important points. I'll write separate posts on points 2 and 4 to be sure that everybody sees them. 1. The rounds are identical, there won't be any new data. The purpose of the rounds is to distribute the scoring over the entire leaderboard phase as it is computationally very intensive. Without the weekly rounds, we would risk to get many submissions in the last week and few submissions before. 3. Yes, NS (number of significant modules) is the main score, it's the number of modules that show significant enrichment in at least one GWAS dataset. --daniel

Questions about challenge structure page is loading…