On Wednesday 06 April 2011 16:18:53 Peter wrote: > Does this sound reasonable (no hits mapped-> no output)? It does. I cannot remember on top of my head whether I changed this behaviour, I think at one point in time MIRA complained if it did not find something to assemble. At least in de-novo I think it did. > If so, could you write some "empty" output files anyway please? Hmmm, would have to check whether easily doable ... or whether I'd have to implement tons of other checks because running the assembly with no overlaps certainly is not something MIRA is prepared for > [..] > Similarly the coverage wiggle file would all be > zeroes, while the contigs would be just the reference sequences. Nope, it would all be ones. MIRA doesn't know "per strain" wig files and the reference sequence always counts in. > Alternatively, how about printing a warning/error message and > giving a non-zero return code? The simplest solution (for me), but then again I'm not certain I like the idea ... while there was nothing to assemble, this is nothing wrong per se. Have to ponder on that. What do others think? B.