You would be better finding papers which have produced the data you're looking for and seeing where it has been uploaded (or if it hasnt, ask them for it).
There is a repository for microbiome data at http://qiita.microbio.me/ which you can download from but I don't know if they have your organism, and it's not the most easy to navigate.
ok if i have understood well the problem is that actually the studies are focused on the Human Microbiome project. and perhaps the lower number of sequences is due to no data for all of the organisms and reflect the few studies. perhaps a solution is test every of these sequence with a BLAST search.