Establish submission strategy

[jvanheld]

We now dispose of motifs for all the datasets, but there are some choices to be made about the best strategy. I am now doing some tests with the leaderboard evaluation tool in order to chose the final one.

Among the possibilities:

• use the primary motifs generated by peak-motifs, or motifs resulting from matrix-clustering applied on these motifs

• on either of these types of motifs select motifs s (primary matrices from peak-motifs or clustered ones), inspected thte matrix-quality graphs in order to identify the most relevant matrices for each dataset

• run on either of these types of motifs (primary matrices from peak-motifs or clustered ones) a optimisation based on a new genetic algorithm named optimize-matrices-GA that I developed with my son. This algorithm significantly increases the AuROC with the datasets that I am using, but I now have to check if it improves the performances on IBIS submission site for leaderboard

I am now submitting leaderboard results with these alternative strategies on the IBIS submission site.

[jvanheld]

Still not sure I understood how to link a submission to a given data type. To be checked on IBIS instructions.