As I understand we submit only 10 locations for 1297 cells, e.g:
1 100, 103, 104, 99, 98, 95, 105, 106, 92, 91, 120
2 3,4,5,6,7,8,9,10,11,12
...
(ref: https://www.synapse.org/#!Synapse:syn15665609/wiki/583251)
I understand the first method of scoring: using the average distance from the "ground truth" (that is the result of the cited paper using 84 genes).
But how is the second score being done? are you using DistMap (https://github.com/rajewsky-lab/distmap) to calculate vISH expression pattern for the 84 genes and then somehow measure the diff to the original vISH?
Created by Dani Livne zurkin Hi no need to calculate vISH, we will just correlate (somehow weighted) the expression of each of the 84 genes across the 1297 positions you submitted to the gene expression using actual positions as calculated using the "ground truth".
thanks
P
Drop files to upload
Challenge scoring: adequate prediction of gene patterns page is loading…