You know, it's funny - there's a couple places where this kind of thing seems to have come up. There's research around crop nutrition levels that shows decreases in nutrient levels as yield per acre goes up, there's research on supplements and vitamins that shows synergistic effects between seemingly unrelated substances, we've seen surprising effects from adding or removing species from an ecosystem. One begins to suspect that the "reductive" method of science - the sort of physics- or mathematics-type "reduce the variables of the problem until we can isolate effects" approach - isn't particularly well-suited to dealing with biological systems. You see it in bioinformatics as well - we've sequenced the genomes for many organisms, and have learned a lot from doing so, but we're also learning the limits of that approach pretty strongly - the organism isn't defined just by its 'code', but its environment; the presence, distribution, and concentration of various chemicals; etc. I suspect as we move more towards the "biological" century here we're going to have to readjust how we approach things to start trying to find those synergistic effects earlier in the process, rather than pull everything down to its constituent parts and then experimenting pairwise with various combos. I get the difficulties in doing that, but I feel like we've repeatedly found the stuff we've discarded as irrelevant to the problem ("things that are not in the visual wavelength the eye perceives") in fact do wind up being relevant (wider full-spectrum light has effects outside the mere spatial perception of objects).