correctly using "--keep-read-assignments" generated files

[AlexandreThibodeauUdM]

Hello all!
Thanks for the new version!

I have tried to make use of the assignment files. My idea is that a lot do not nesseraly know the species that are extremely hard to classified, like for exemple veillonella species. So my idea is to use the assignment files so that a user can go in themand llok up to a certain species and have a feeling on how sure they are that this is the right species, instead of going all-in in a useless discussion.

So I have used R to generate this image. The gray bars are the number of sequences that returned a probability for said species and the red dot the median probability.

What do you think. I am completely wrong or is this usefull ?

[kdcurry]

I do like this idea, but I think a slightly different presentation would be more useful. Specifically for the "number of sequences". Let's say species X is included in the probability of 100 sequences. If each of those probabilities were all 99% that would be quite different than if they were all 1% - and the reported abundance would be quite different as well. Therefore, I think instead of "number of sequences that returned a probability" the more appropriate metric would be something like "emu final estimated counts", or a least something that takes the probabilities into account rather than simply the number of sequences that included it in their distribution.

[AlexandreThibodeauUdM]

Sorry it took me time to answer. I see what you mean. Maybe I should just used the final estimated count (I use -keep-count in Emu) for all species. My goal was to show what Emu "had chosen" in the end.