Testing representativity

Just a quick note here. We had run some tests on mixtures of plants to see how representative our DNA data are of the relative proportions of the biomass of different three species that were mixed together.

We put together 10 mixtures of leaves of 4 plant species. 2 were alfalfa, 1 was lotus, 1 was Onobrychis. Mixtures were either 50-50 or 90-10.

We then compared the expected % of mixture with actual % of reads from each species.

In short, about 85% of variation in actual % of reads could be explained by % in mixture and the species identity. Those plants that were higher in the mixture were higher in the reads. Alfalfa read a higher percentage than was expected, but this is likely due to higher protein concentrations than the other species.

We could do more sophisticated analyses and other tests (results from feeding trials are forthcoming), but this is a pretty good start. It supports the idea that the relative percentage of reads should indicate relative intake of protein of different plant species.