zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Commit Graph

  • 8dca236958 Base-packed reader cleanup. hanna 2009-10-15 01:26:23 +0000
  • 316b30ee56 On the road to human: make sure the suffix array will fit in a Java array. hanna 2009-10-14 21:45:35 +0000
  • e740e7a7ce Because walkers call UG's map function, we need to move the actual writing out to UG's reduce function. ebanks 2009-10-14 20:49:26 +0000
  • 825e6c7a4d added calculation for bases over 2x,10x,20x,30x plus gene name kcibul 2009-10-14 20:32:26 +0000
  • 727b69fce0 catch null output destinations earlier aaron 2009-10-14 20:07:15 +0000
  • 1f66738c8e Fix a hashing function bug. Ignore reads with non-reference bases in the pileup. chartl 2009-10-14 19:41:26 +0000
  • 72c34f11dd Bug fixing for BWA output formats. hanna 2009-10-14 19:32:22 +0000
  • 60183229ab the oldest java mistake in the book... aaron 2009-10-14 19:32:13 +0000
  • 52d2e0ca07 All walkers now use read.getReadGroup() ebanks 2009-10-14 19:27:40 +0000
  • 0a09fa4d5c Rename to distinguish this transition table calculator from the scala version. chartl 2009-10-14 18:52:21 +0000
  • 1d055011bd Getting rid of this so I can rename it without the world blowing up. chartl 2009-10-14 18:45:11 +0000
  • eb90e5c4d7 changes to VCF output, and updated MD5's in the integration tests aaron 2009-10-14 18:42:48 +0000
  • 89771fef05 -Use read.getReadGroup() -Add another filter for read groups for Chris ebanks 2009-10-14 18:08:32 +0000
  • 311ab8da5a A helper class to create the masks for the sequenom design maker. This project is now officially done. ebanks 2009-10-14 17:28:51 +0000
  • 3553fc9ec0 Preparing for human -- support bwa output files directly rather than relying on a custom fixed sa interval. hanna 2009-10-14 17:17:46 +0000
  • d89bc2c796 This class no longer outputs in sequenom format ebanks 2009-10-14 17:16:36 +0000
  • 0c95d6906f Merge both versions of the Sequenom assay design maker: use Jared's base code and add in indels. [Jared, this still emits the same output for SNPs as your original version) Remove all sequenom stuff from the FastaAlternateReferenceMaker so it can just concentrate on making alternate references... ebanks 2009-10-14 17:11:45 +0000
  • 49af5269e5 Jared: feel free to change or revert, but until we move over to UG version... Only print out positions with at least one non-ref call ebanks 2009-10-14 17:08:57 +0000
  • f5a2e6dd50 Fix! chartl 2009-10-14 16:15:20 +0000
  • f2886d88e0 We now emit genotype calls ebanks 2009-10-14 02:49:56 +0000
  • 1b214c0de5 Fixed logic: throw exception if contigs are NOT equal ebanks 2009-10-14 02:48:44 +0000
  • aeca14d052 On our side of 5CC, we spell multi M-U-L-T-I. ebanks 2009-10-14 01:41:25 +0000
  • c9c8fd1fef Added the discovery LOD score to the meta data ebanks 2009-10-14 01:24:06 +0000
  • 0c06bf9dbc Explicitly set output to GELI now that default is VCF ebanks 2009-10-13 22:12:03 +0000
  • a76fac4687 Cleanup existing speedups. Minor performance improvements. hanna 2009-10-13 21:51:18 +0000
  • 837ae1d33a Optimization: from 22k reads/min - 30k reads/min. hanna 2009-10-13 20:59:29 +0000
  • 96b8499a31 Remodeled version of the UnifiedGenotyper. We currently get identical lods and slods as MultiSampleCaller (except slods for ref calls, as I discussed with Jared) and are a bit faster in my few test cases. Single-sample mode still emulates SSG. The remaining to do items: 1. more testing still needed 2. we currently only output lods/slods, but I need to emit actual calls 3. stubs are in place for Mark's proposed version of the EM calculation and now I need to add the actual code. More check-ins coming soon... ebanks 2009-10-13 20:27:01 +0000
  • b28446acac Multi-sample calls now have associated meta-data (SLOD, allele freq), which wil l soon actually be used... ebanks 2009-10-13 20:08:43 +0000
  • db642fd08b Optimization: from 10k reads/sec - 22k reads/sec.. hanna 2009-10-13 18:07:15 +0000
  • 77499e35ac fixes for GSA-199: Need easier way to write binary outputs to standard output. GLF and VCF now have stream constructors, and can get dumped to standard out. aaron 2009-10-13 15:50:20 +0000
  • f37564e63a Our BWA is now looking at roughly the same number of candidate alignments as BWA/C. Performance is now at 11k reads / min, still a long way from BWA/C. hanna 2009-10-13 15:50:04 +0000
  • 8d0e057d83 I got bored today and decided to write the confusion matrix calculator. At present it is untested. I'm submitting it to subversion to make sure I have previous revision to revert back to. chartl 2009-10-13 02:55:29 +0000
  • be92a1e603 Don't try to close if the lazy initialize hasn't triggered ebanks 2009-10-13 01:20:25 +0000
  • ec83bc6ec5 This somehow didn't make it into subversion the last time. chartl 2009-10-12 21:11:13 +0000
  • ecbb11e017 Modified PowerBelowFrequency to ignore reads below a user-defined mapping quality. Request from Jason Flannick. chartl 2009-10-12 20:59:24 +0000
  • ec68ae3bc5 Added a filter that will split the read set by a threshold of mapping quality (Request from Jason Flannick) chartl 2009-10-12 20:58:37 +0000
  • 0d73fe69e7 Recalibrator by NQS. Had this puppy running all afternoon. Thing had got through 100,000,000 reads before I decided to delete my sting tree. *sigh*, a little more delay. chartl 2009-10-12 20:55:02 +0000
  • ee0afba0af Recalibration stuff... chartl 2009-10-12 20:51:39 +0000
  • caf689821f added method to get normalized posteriors ebanks 2009-10-12 02:33:22 +0000
  • cf7a26759d -use the getReadGroup() function that was added to picard for us -clean up some include lines ebanks 2009-10-12 01:39:32 +0000
  • d844d1c496 SAMFileWriters specified as command-line arguments were sometimes incorrectly altering the default short name. Make sure short name is not specified if shortName is not specified but fullName is. hanna 2009-10-09 19:16:46 +0000
  • da084357db Fixed minor typo in output message. hanna 2009-10-09 18:56:54 +0000
  • 62c484b57a Fixes for GSA-201, where enumerated types in command line arguments had to be defined as all uppercase for the system to work. aaron 2009-10-09 18:11:32 +0000
  • 32d55eb2ff Fix issue Eric was seeing with java.lang.Error in unmap0. hanna 2009-10-09 17:46:56 +0000
  • 9f3482ef11 VCF is both a multi- and single- sample format, so we shouldn't be throwing an exception when used for SS ebanks 2009-10-09 17:43:26 +0000
  • d9f5a314ac avoid an out of memory error by no putting more than 5000 reads in the cache. on pilot1 at least those are crazy loci anyway. jmaguire 2009-10-09 14:56:55 +0000
  • f4b6afb42c JVM issue id 5092131 (http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=5092131) was causing OOM issues with the new mmapping fasta file reader during large jobs. Temporarily reverting the reader until a workaround can be found. hanna 2009-10-09 04:45:46 +0000
  • 6d7f4481e4 Changed traversal type slightly chartl 2009-10-09 04:11:48 +0000
  • a9f3d46fa8 Your time has come, SSG. Fare thee well. ebanks 2009-10-08 20:27:56 +0000
  • 8fdb8922b8 now output in the exact format that works with sequenom software. jmaguire 2009-10-08 20:06:27 +0000
  • 98e3a0bf1a VCF can now be emitted from SSG. The basic's are there (the genotype, read depth, our error estimate), but more fields need to be added for each record as nessasary. aaron 2009-10-08 19:50:04 +0000
  • 95f24d671d Fixed 'visualization' of reads that didn't match bwa's alignments exactly. hanna 2009-10-08 19:45:30 +0000
  • 29ad6cd876 Made redundant by BCMMarkDupes kiran 2009-10-08 18:47:20 +0000
  • 94d82d1915 Matthew Bainbridge's duplicate removal utility for 454 data. This code should eventually be moved into a read walker. For now, it's being introduced into the repository as-is (well, with one minor change to make the handling of command-line arguments a little more straightforward). kiran 2009-10-08 18:32:37 +0000
  • 8461cc3a22 /tmp is failing... We should be using hptmp anyways. ebanks 2009-10-08 18:13:49 +0000
  • 15bf014e0b logger.info -> logger.debug (don't want to risk filling up my log on genome-wide calls) ebanks 2009-10-08 17:53:11 +0000
  • df8ea8f437 UG integration test. This was the old SSG test with MD5s updated. I'll need to add some multi-sample tests in a bit... ebanks 2009-10-08 17:43:58 +0000
  • b39a307a04 SSG -> UG ebanks 2009-10-08 17:40:09 +0000
  • 008455915a One way of making the integration test stop failing is to remove it... [waiting for Matt to cringe...] ebanks 2009-10-08 17:08:41 +0000
  • f89a89ffe3 Use of AlleleFrequency as an input to PowerAndCoverage is deprecated by the new walker. Reverting to the standard "power at 1 allele" calculation. chartl 2009-10-08 16:07:45 +0000
  • ae05f5c7ad Fixin the header. chartl 2009-10-08 15:49:28 +0000
  • 11ff1e09b8 A new power walker for the user to feed in a number of alleles. Call that number k. Output is: chartl 2009-10-08 15:35:35 +0000
  • 04fe50cadd *** We no longer have a separate model for the single-sample case. *** ebanks 2009-10-08 14:08:57 +0000
  • 32128e093a misc. changes to get the numbers back to the baseline while keeping the speedup. jmaguire 2009-10-08 12:27:07 +0000
  • d38a0d04b9 fix a snp mask offset error. jmaguire 2009-10-08 12:25:40 +0000
  • 829e99413b Rescores a variant after removing duplicates (defined very strictly as reads with the same start points). kiran 2009-10-08 03:07:36 +0000
  • fcb6a992c8 Switched IndexedFastaSequenceFile over to use memory mapping to load data rather than the loop-with-small block size. Performance improvements in loading refs are extreme; segments can be loaded in <1ms. chr1 in its entirety can be loaded in 1.5sec (down from 30sec). hanna 2009-10-08 00:07:15 +0000
  • 02d2492d68 Simple tool for picking sequenom probes for SNPs. Can be extended to indels if necessary. jmaguire 2009-10-07 23:46:41 +0000
  • 1905b5defa Hash by chromosome for now to reduce memory. This is a temporary solution until we decide how to reture the Injector for good. Also, with Picard's latest changes, we need to make sure we don't double-close the sam writer. ebanks 2009-10-07 20:06:25 +0000
  • f9a1598d75 Reformatting ebanks 2009-10-07 20:03:34 +0000
  • 203c626fc2 A wrapper around the GenotypeLikelihoods class for the UnifiedGenotyper. This wrapper incorporates both strand-based likelihoods and a combined likelihoods over both strands. ebanks 2009-10-07 19:57:37 +0000
  • 5bdcc2b4dc Included HLA class 2 genes in CreatePedFileWalker sjia 2009-10-07 18:46:51 +0000
  • 8f896b734f Included HLA class 2 genes in CreatePedFileWalker sjia 2009-10-07 18:28:01 +0000
  • f9a0eefe4b GELI_BINARY is now functional, and can be used as a variant type in SSG (-vf=GELI_BINARY). Also fixed the max mapping quality column in both GELI output formats, we haven't been correctly outputing up until now. aaron 2009-10-07 18:20:34 +0000
  • 225b9bccc1 Modifications to NQSClusteredZScoreWalker to output empirical mismatch rates on bins by both Z-score and reported Q-score, rather than averaging over all Q-score bins for each Z-score. chartl 2009-10-07 13:45:12 +0000
  • 8dd0924b37 Minor performance improvements to VariantEval -- now all of the CPU time is spent dealing with the ROD system... depristo 2009-10-06 23:40:30 +0000
  • 4554ca1b28 more cleanup, depecaited the old genotype, corrected SNPCallsFromGenotypes' imports and two other classes that depend on it. aaron 2009-10-06 19:09:27 +0000
  • 3aec76136f Removing the AllelicVariant interface, which is replaced by the Variation interface. aaron 2009-10-06 17:44:24 +0000
  • c1e1d910cb simple monitor for watching pilot 1 call progress depristo 2009-10-06 13:04:53 +0000
  • 1bd0c3c145 variant eval allows non Variation rod objects depristo 2009-10-06 13:04:26 +0000
  • 66fc8ea444 GSA-182: Adding support for BED interval files. aaron 2009-10-06 02:45:31 +0000
  • aec83b401d SSG multithreading doesn't play well with some I/O changes made since I last svn up'd. Reverting until I can find the reason. hanna 2009-10-05 19:48:57 +0000
  • 8a503c86b6 Code supporting SSG proof-of-concept shared memory parallelism. hanna 2009-10-05 18:56:16 +0000
  • fb619bd593 -Refactoring: make GenotypeCalculationModel constructors empty so that they don't have to be updated every time we add a new parameter; instead put that logic in the super class's initialize method (making everything protected so that only the factory can access them) -Adding initial version of Multi-sample calculation model. This still needs much work: it needs to be cleaned up and finished. Right now, it (purposely) throws a RuntimeException after completing the EM loop. ebanks 2009-10-05 18:10:36 +0000
  • 98076db6b4 Modified CreatePedFileWalker to output PED file given HLA allele names sjia 2009-10-05 03:06:42 +0000
  • 56bc4fa21a Fixed bug where not all alignments were returned if read aligned to multiple locations. Enhanced test suite to validate all alignments. hanna 2009-10-04 18:20:20 +0000
  • 05aa928e3e Fix off-by-number-of-deletions issue with negative strand reads. Improved performance by factor of 2.5x. hanna 2009-10-03 21:55:18 +0000
  • 7605ee500c Idiocy! All tests were being disabled because I forgot the instanceof chartl 2009-10-02 20:04:56 +0000
  • 88d0890cc3 Made PooledGenotypeConcordance a standard test in VariantEval chartl 2009-10-02 20:03:31 +0000
  • 7fc4472e6d A big fix for MergingSamRecordIterator, where we weren't correctly handling the comparisons of SAMRecords correctly (we weren't applying the new reference index first, so sometimes the MT contig would be ID 23, sometimes 24 in different records). aaron 2009-10-02 19:35:35 +0000
  • 68cb2ee54b Tweaks to parameters for NQS analysis walkers; change to PowerAndCoverage for Jason Flannick (can input the number of alleles to compute power for - i.e. doubletons, tripletons; rather than statically checking singletons. chartl 2009-10-02 19:11:27 +0000
  • 7249fade05 updated ebanks 2009-10-02 18:10:34 +0000
  • 53a4bd7f51 A better understanding of what's going on means no need for clearing the cache ebanks 2009-10-02 18:07:46 +0000
  • e885cc4b21 changes for corrected GLF likelihood output, along with better tests aaron 2009-10-01 20:45:05 +0000
  • 2309d19f6f Bug fix from Michael Ross: mark second read in sequence as second of pair. hanna 2009-10-01 14:34:36 +0000
  • 48b478f751 epic fail! aaron 2009-09-30 20:49:42 +0000
  • 2e4949c4d6 Rev'ing Picard, which includes the update to get all the reads in the query region (GSA-173). With it come a bunch of fixes, including retiring the FourBaseRecaller code, and updated md5 for some walker tests. aaron 2009-09-30 20:37:59 +0000
  • 303972aa4b Yup, I broke the build... ebanks 2009-09-30 20:20:43 +0000
  • 841d25cc44 Added ability to set the priors after construction (and requiring a flushing of the likelihoods cache) ebanks 2009-09-30 19:55:49 +0000
  • 665951f9f0 Support negative strand alignments. hanna 2009-09-30 18:10:26 +0000