We currently are limited to remote sensing, which pretty much boils down to optical and near-IR spectroscopy, so we're looking at upper atmosphere molecules.But how far out of chemical equilibrium could you get without life? It doesn't seem to happen too much in our solar system. Not only, as Lovelock pointed out, does life have to use its environment to obtain resources and dump wastes, living things tremendously increase the surface area available for various interesting reactions. Maybe each planet should be compared to others in the same solar system.
Well, first thing to look for are atmospheres out of chemical equilibrium; but that in itself is not sufficient as a biosignature, since there is significant flow of free energy from the star, and photolytic chemistry can produce chemical disequilibrium, including molecular oxygen.
Now, 'we' think we understand photolytic gas chemistry, and bottom line is that if you see oxygen (or ozone proxy) AND methane, in the presence of water (and we expect some carbon dioxide as well), then we have carbon based photosynthetic life... or some funky chemistry we didn't predict.OK, so that should be corrected to read, "for a large fraction of the Earth's history, there was quite a bit of photosynthesis and it was even oxygenic, but the oxygen was getting mopped up by iron until the iron ran out". Still, it's a valid point and a fascinating question. What set of characteristics would allow us to distinguish a planet with life from one without it?
BUT, for a large fraction of the Earth's history, there was no significant photosynthesis, and life utilized anaerobic metabolism, but it was still life.
So, what is that signature: we're talking methanogens, free-hydrogen metabolising critters, others living on sulphur compounds or metal ions. We don't even know what the Earth's atmosphere was like at different stages in the archaean, much less what the range of possible atmospheres was which was consistent with pre-aerobic life.