There are number of methods to screen your data. You could try multivariate methods like principal component analysis, for example. Best performing strains will pile up on one axis of the component. Regression trees and random forests are also popular methods for screening.