Hi ! I modified yesterday the algorithm that find the peak. New code is simplier and smaller. No smoothing involved.
Although I have 35 unit tests to test this 60-lines function (CoverageDistribution::CoverageDistribution), it can probably be improved more. Can you send me (pastebin) the CoverageDistribution.txt file for each of your runs. I will add them to my unit tests. Also, do you have a place where I could upload you data-for-unit-tests and data-for-system-tests so that you can run tests ? 22G data-for-system-tests 6.8M data-for-unit-tests There is a README in system-tests and unit-tests too ! Thank you ! > ________________________________________ > De : David Eccles (gringer) [[email protected]] > Date d'envoi : 21 juillet 2011 06:13 > À : [email protected] > Objet : [Denovoassembler-users] No observed peak in coverage distribution > (454 test data) > > I tried to a short test run with "real" data using the most recent > sebhtml/ray git, and got the following error: > Rank 0 has 4463742 k-mers (completed) > > Rank 0: the minimum coverage is 1 > Rank 0: the peak coverage is 1 > Rank 0: Assembler panic: no peak observed in the k-mer coverage > distribution. > Rank 0: to deal with the sequencing error rate, try to lower the k-mer > length (-k) > Rank 0: sent 241979 messages, received 241978 messages. > > I get the same outcome with my 5k phix assembly test: > $ ../../../sebgit/ray/code/Ray -p tests/phix/phix_5k_1.fasta > tests/phix/phix_5k_2.fasta > ... > > Rank 0 has 10466 k-mers (completed) > > Rank 0: the minimum coverage is 1 > Rank 0: the peak coverage is 1 > Rank 0: Assembler panic: no peak observed in the k-mer coverage > distribution. > Rank 0: to deal with the sequencing error rate, try to lower the k-mer > length (-k) > Rank 0: sent 201223 messages, received 201222 messages. > > When this happened previously, I managed to get around it by re-writing > the smoothing function > (https://github.com/gringer/ray/commit/f048203e571d4c7a267dace27fec063cfe2059dd), > but you're no longer using a separate smoothing function in your > coverage analysis code. Why are these assemblies failing? > > The data I'm working from is here: > http://user.interface.org.nz/~gringer/hacking/planaria_454.fasta [about 7MB] > http://user.interface.org.nz/~gringer/hacking/phix_5k_1.fasta [about 300kB] > http://user.interface.org.nz/~gringer/hacking/phix_5k_2.fasta [about 300kB] > > Cheers, > David > > ------------------------------------------------------------------------------ > 5 Ways to Improve & Secure Unified Communications > Unified Communications promises greater efficiencies for business. UC can > improve internal communications as well as offer faster, more efficient ways > to interact with customers and streamline customer service. Learn more! > http://www.accelacomm.com/jaw/sfnl/114/51426253/ > _______________________________________________ > Denovoassembler-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/denovoassembler-users > Sébastien ------------------------------------------------------------------------------ 5 Ways to Improve & Secure Unified Communications Unified Communications promises greater efficiencies for business. UC can improve internal communications as well as offer faster, more efficient ways to interact with customers and streamline customer service. Learn more! http://www.accelacomm.com/jaw/sfnl/114/51426253/ _______________________________________________ Denovoassembler-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/denovoassembler-users
