Conflicting taxonomic assignments of gene calls

[Caroline Fortunato]

Hello,

I have run the same dataset twice through IMG, once in 2014 and once a few days ago and I've noticed that the taxonomic assignments reported using the latest IMG pipeline version are drastically different compared to the same datasets run using an earlier IMG pipeline version. Independent 16S analysis, mapping to genomes, and BLAST calls correspond to the older taxonomic assignments on the data. The current taxonomic assignments do not match other results from our analyses. If someone could explain how the databases IMG is using for the pipeline have changed that would be helpful. I am curious to understand what genomes are/are not being included and why. 

Thank you,

Caroline Fortunato