Thank you all! You helped me a lot. Cheers Dani. From: bach@xxxxxxxxxxxx Subject: [mira_talk] Re: Too many redundant contigs Date: Fri, 13 Jan 2012 21:02:12 +0100 To: mira_talk@xxxxxxxxxxxxx On Jan 13, 2012, at 14:19 , Sven Klages wrote:2012/1/13 Remo Sanges <noncoding@xxxxxxxxx> What I find useful when I want to reduce redundancy is to use the output of mira (.unpadded.fasta + .unpadded.fasta.qual) as input for cap3. Well, ok, why not. Personally I'd not go that way. But IMHO it is still quite important, to find out what "redundant " really means. Just parts of contigs with homologies? Or some artefacts from library construction in the sequencing library (we used a kind of custom approach years ago, I needed to individually clip these remainders)? Both approaches have their pros and cons. At work I do have some RNASeq libraries to analyse and the biologists are sometimes appalled by the number of mutations present in the data. Of course, each and every one must be analysed in-depth because it could be important *sigh* When she (roughly) knows why these contigs are not merged, she can start to alter assembly strategy accordingly. Or simply continue with the data as it is. Doing a condensing too early is sometimes a bit dangerous :-) B.