I have 80 independent sets of genes with 6000 genes in each. Gene expression values are ranked and scaled from 0 to 1. So, a set looks like this: Gene1 1 Gene2 0.98 Gene3 0.85 ... Gene5998 0.002 Gene5999 0.0015 Gene6000 0.00001 Each set has 6000 of the same genes, but of course, the expression values are different in each set. However, I should have several couples of sets that are very similar. By similar, I mean that the majority of genes should have similar values in two sets. What will be the best way to find those similar sets? And how to show that similarity? One way that I was thinking is to take set A and set B, remove all genes with expression lower than some threshold (like 0.01) in both sets, and find Jaccard similarity between gene sets. But it is hard to set a threshold in this case. More details as requested: Those 80 gene sets were obtained from two different datasets. The goal is to see if some of the 40 gene sets from dataset 1 are similar to some of the 40 sets from dataset 2.

Full article content could not be extracted automatically. Read the original below.