Genes mentioned along with word in Figure or Table category

Right now, this seems to be the most promising approach.

'Over-association' will come from things like markers.

'Under-association' will come from lists of genes in figures or tables, but not mentioned in the text in combination with a word from the figures and table category.