Several years ago I was one of the co-authors of the program RNAmmer. This program was created to find ribosomal RNAs in genomic sequences. The program uses Hidden Markov Models to find sequences that are sufficiently similar to other rRNAs within that taxonomical kingdom. Our primary focus was on bacterial rRNAs, mostly due to the lack of data for eukaryotic and archaeal rRNAs, which made making models for these more difficult. One of the consequences of this is that we did not make models for the eukaryotic 5.8S rRNA.
Due to personnel issues (i.e. none of us being employed in steady positions which allowed us to work on this program), we have not been able to keep on top of the program as we would have wanted to. However, we are now in the process of gearing up to a version 2. This version will include new models, potentially also for the eukaryotic 5.8S rRNA, and will also be compatible with HMMER 3. It will also be released under an open license, making it easier for people to use it and to contribute to its further development.
This is where you can find the original paper.
Link to the current version of the software.
Please note: there are several known hickups with the existing program. We will be releasing a new version shortly which will make it easier to install and use. In the mean time, I have written a blog post describing what needs to be done to get RNAmmer 1.2 up and running.
I recently tried using the .hmm files from RNAmmer 1.2 with HHMer 3.1b2, specifically nhmmer, and they seem to work straight out of the box, and fast enough that I can use the full .hmm models against a metagenome assembly (> 500MB) without screening by the abbreviated models first. Have you tested this approach in your own experiments?
This sounds great! I will have to check that. Thanks for letting me know!