Hi,
At the moment, Prodigal does not multithread or have any parallel version. A few users have requested this, though, so maybe someday. The metagenomic version is "embarrassingly parallel", however, so you can split up the input file and start many prodigals running on the different pieces.
What I would probably do, rather than make the prodigal C code multithreaded, is just make a multithreaded python wrapper that split the sequences into sub-files (call it "metaprodigal") and spawned one prodigal per requested thread, then collected the results. Is this something that would interest the rest of the discussion group?
regards,
doug