I have a fasta file of about 200,000 RNA sequences and a server holding a local copy of Rfam. For each sequence I want to know the type of RNA it is most related to and ultimately retrieve statistics on the proportion of sequences that are most probably rRNAs, tRNAs, real sRNAs etc.
I'm sure this kind of thing has been done before plenty of times but I can't find much information on it myself. I was going to use BLAST for the job, retrieving the top scoring result from each output report using perl. My question is, is BLAST really capable of handling the local alignment of many very short (16 - 50nt) sequences in a reasonable amount of time?