That's a good point. A good metric that Ray could produce to start with would be the number of pairs (including mates) with:
1. both ends within a contig; 2. one end on one contig and the other end on another contig 3. one end on one contig and the other not mapped 4. both ends not mapped You suggest that a sizable part of the pairs (including mates) arein 3. and 4. when using a k-mer length of 61-91. That's likely. I think it is probably the case as mate pairs usually include also an adapter too, and that consume previous space in the sequences. For the time being, I believe that "use another scaffolder" is your best bet. Speaking of scaffolders, I will soon (hopefully) fix the speed issue for scaffolding of large genomes due to repeated k-mers [1]. -- [1] https://github.com/sebhtml/ray/issues/91 On 12/13/2012 04:20 PM, Louis Letourneau wrote: > Most big genome assemblies use pairs and mates to build their genomes. > > Pairs a great but short > Mates are long and have pairs mixed-in + chimeric reads (in the sens that the > jumping juncting is IN a read) > > If you have good sequencing you want want to use a big kmer to avoid > repeating regions with the pairs, but you want short kmers with the mates > because of the extra errors (with smaller kmers you could save the chimeric > if your kmer ends before the junction). > > something like 61-91 for pairs > 21-51 with mates. > > Right now either you run everything in the middle...51-61 > Use another scaffolder > Rerun your assembly in ray again with mates added (not sure if this would > work) > > Does this make sens or I'm missing a feature of ray that will work even with > bad reads? > > Louis ------------------------------------------------------------------------------ LogMeIn Rescue: Anywhere, Anytime Remote support for IT. Free Trial Remotely access PCs and mobile devices and provide instant support Improve your efficiency, and focus on delivering more value-add services Discover what IT Professionals Know. Rescue delivers http://p.sf.net/sfu/logmein_12329d2d _______________________________________________ Denovoassembler-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/denovoassembler-users
