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

Commit Graph

  • 990af3f76e Will now work with simplest tabular format - genotype string ("+ACTT") does not have to be followed by ':' asivache 2010-02-01 15:40:01 +0000
  • e0808e6c37 Moved old EM model to archive ebanks 2010-02-01 02:55:32 +0000
  • 64fc76e4bf Added an option to AnalyzeCovariates to set the max value of the histograms to make them easier to directly compare. rpoplin 2010-01-31 23:13:57 +0000
  • f6da57dc79 1. For Matt: JIRA GSA-270. Other walkers needing to call into the Unified Genotyper now use static methods (e.g. runGenotyper()) instead of calling initialize and map. ebanks 2010-01-31 21:14:57 +0000
  • e964660df3 snpSelector now supports min and max q scores. depristo 2010-01-31 19:38:34 +0000
  • 7b3c34d210 keeping a backup depristo 2010-01-31 15:36:25 +0000
  • ce9d3dcefb Removing deprecated version of indel genotyper (putting it in archive in case we need to reproduce original 1KG indel calls for some reason). ebanks 2010-01-31 14:05:36 +0000
  • 3d45457595 VariantEval2 test framework implemented; Kiran is experimenting with the system. Not for use by anyone else. VariantContext appears to work well; I'll release it next week for general use following docs of the functions. Removing newvarianteval and other classes to avoid any future confusion. Update to TraverseLoci and RodLocusView to simplify a few functions and to correct some minor errors. All tests pass without modification. depristo 2010-01-30 20:51:24 +0000
  • 236764b249 Major (and useful) changes to MultiSampleConcordance: chartl 2010-01-30 01:18:31 +0000
  • ea7e737441 Two new annotations: jmaguire 2010-01-29 23:23:00 +0000
  • 97f60dbc4b Moving stuff around. ( core;playground ) ----> ( oneoffs ). I've been a bad boy, sullying the core codebase. chartl 2010-01-29 22:50:03 +0000
  • 16da5011c0 Added a new option for indicating the mean number of variants on the AnalyzeAnnotations plots. This way one can say, for example, filtering at this point will keep 75 percent of all the variants. rpoplin 2010-01-29 21:58:31 +0000
  • 668c7da33d Bug fix in custom override of queryOverlapping. hanna 2010-01-29 21:35:59 +0000
  • c6cc844e55 Added -name argument to AnalyzeAnnotations that allows one to specify the name of the annotation to be used on the plots. Instead of seeing AB and DP, one can add -name AB,AlleleBalance -name DP,Depth rpoplin 2010-01-29 20:48:53 +0000
  • 62a80f2b6f fixed out of date tests. Also, tests uncovered a subtle bug in new implementation that was also fixed depristo 2010-01-29 20:03:48 +0000
  • 4f29a1d4f6 AnalyzeAnnotations now plots true positive rate instead of percentage of variants found in the truth set. Committing GCContentCovariate to help people experiment with correcting the pilot3/Kristian base calling error mode in slx. rpoplin 2010-01-29 20:01:56 +0000
  • ac2a207b0b added a wrapper exception for anything that goes wrong in VCF parsing; this way the problematic file line is emitted, no matter what happens. Makes debugging a lot easier, especially in large files. aaron 2010-01-29 19:58:51 +0000
  • e7f5c93fe5 Cleaning up the inheritance hierarchy from the previous commit. hanna 2010-01-29 19:13:36 +0000
  • 88495a39d4 better formating depristo 2010-01-29 15:38:21 +0000
  • 1993472b38 Just like VariantFiltration but lets you match info fields out of the VCF instead of annotating them. depristo 2010-01-29 15:38:03 +0000
  • 0a7426c29c Computes SNP density over the genome. Doesn't work with intervals depristo 2010-01-29 15:36:49 +0000
  • 9decd20f46 Fix to priors to allow lower het values for mouse guys; no intergration test changes depristo 2010-01-29 15:36:12 +0000
  • d57a86ad41 Not nearly as badass as it looks. The problem I mentioned yesterday with "bleeding in" of samples comes from VCFUtils and SampleUtils looking for all VCF-class RODs in the tracker, and stealing the name from them. I have introduced a new HapmapVCF - type rod for use when you want to protect your VCF header from being infected by the samples in a bound hapmap VCF. Changes are as follows: chartl 2010-01-29 15:19:50 +0000
  • 5aaf4e6434 VariantFiltration now accepts any number of --name --filter expressions, and annotates the VCF file with each name that matches. Very useful depristo 2010-01-29 12:13:08 +0000
  • 01e73fc39e Yuck - Picard's SAMRecord Comparator only deals with mapped reads. Adding an extended version that works for all reads. After adding some more minor changes to the new realigner it now gets the same exact results as the original version - except that sometimes it doesn't clean when it shouldn't! More testing coming. ebanks 2010-01-29 07:49:47 +0000
  • 3d922a019f Basic support for very simple index-driven locus traversals. Interface has been changed to support batched intervals in a single shard, but intervals are not yet compressed into a single shard. hanna 2010-01-29 03:14:26 +0000
  • 4810e9c9cd And now the DOCS! asivache 2010-01-28 23:21:33 +0000
  • 40262e2070 Now calls single-sample indels too, with all the V2 level stats and bells. This officialy obsoletes IndelGenotyperWalker (V1). In addition, the alignments spanning beyond the contig end are now completely ignored (with a user warning), this applies to both single-sample and paired (somatic) calls. You just wait, Eric, I'll get you the docs with the next commit! asivache 2010-01-28 22:28:02 +0000
  • 0fb032a436 Quick script that changes "chr#" to "#" and "chrM" to "MT" and moves mitochondria to the end of the vcf; in accordance with the 1KG reference. chartl 2010-01-28 21:59:33 +0000
  • 79c4cc1db7 AnalyzeAnnotations now breaks out titv by calls in hapmap and also plots true positive rates. Any RODs passed in whose name starts with 'truth' is considered to be the truth set. rpoplin 2010-01-28 21:41:23 +0000
  • 7a10c40fb3 Much clearer (and, like, not totally incorrect) implementation of isNovel chartl 2010-01-28 21:16:21 +0000
  • 8de6a8d246 Lots of changes; all to do something relatively minor. chartl 2010-01-28 21:06:56 +0000
  • 6f11fe442a Sync with Andrey's changes ebanks 2010-01-28 20:49:38 +0000
  • db429e1096 Some alt consenses may have cigar string starting with an insertion. Not a bug, strictly speaking, since the cleaner had been detecting this and crashing deliberately. Now it knows how to deal with this special case though. Also, uppercase the ref before using it in SW aligner! asivache 2010-01-28 18:53:02 +0000
  • 956b570c8e V5 improvements to VariantContext. Now fully supports genotypes. Filtering enabled. Significant tests throughout system. Support for rebuilding variant contexts from subsets of genotypes. Some code cleanup around repository depristo 2010-01-28 18:37:17 +0000
  • 9876645a5d Now drives the walker by reference, not by reads, so we see even loci with no reads. This allows us to accurately calculate the true total callable area depristo 2010-01-28 11:12:46 +0000
  • 1dd9996f3a New realigner now completely uses bytes, plus misc fixes. Still not ready for use. ebanks 2010-01-28 04:17:20 +0000
  • f6bca7873c V3 of VariantContext. Support for Genotypes and NO_CALL alleles. QUAL fields fully implemented. Can parse VCF records and dbSNP. More complete validation. Detailed testing routines for VariantContext and Allele. depristo 2010-01-28 04:10:16 +0000
  • 4642a1ae59 Give the symlink on the webserver a name that would be a bit more obvious. hanna 2010-01-28 00:40:20 +0000
  • 23fc9737b4 Added the ability to filter out variant (not truth) calls based on read depth. Using -NLD 5 will not update concordant counts for calls with 0, 1, 2, 3, or 4 reads supporting them. Not to be used with VCF files that do not have DP in the format field. chartl 2010-01-27 23:28:04 +0000
  • 1b9184a1c7 Added a multisample concordance walker which takes the place of the VCF python library I've been using. Takes a truth VCF and a variant VCF and outputs A TSV that looks like this: chartl 2010-01-27 22:59:17 +0000
  • eccf40b17d Tweak the layout of the resulting .tar.bz2 a bit. hanna 2010-01-27 22:07:37 +0000
  • bd11060e72 Ups, I did it again. Fixing the bug introduced in a previous commit: use correct length of the indel event. asivache 2010-01-27 21:51:54 +0000
  • fddca032bb Initial commit of v2.0 of the cleaner. DO NOT USE. (this means you, Chris) ebanks 2010-01-27 21:36:42 +0000
  • b8ae083d1b AnalyzeAnnotations creates a plot of dbsnp rate as a function of the annotations. rpoplin 2010-01-27 21:08:33 +0000
  • 668defc841 Fix the dreaded 'can't use fully qualified symlinks on the ftp server' gotcha. hanna 2010-01-27 20:55:48 +0000
  • dbd3d4473e Making a copy here before changing the live copy ebanks 2010-01-27 20:44:15 +0000
  • 3999a8d2c8 IntelliJ no longer complains that my methods are too complex to analyze. rpoplin 2010-01-27 20:12:13 +0000
  • fc4285f9fd AnalyzeAnnotations seems to be popular so I've rewritten the guts to be easier to extend and maintain. rpoplin 2010-01-27 19:30:31 +0000
  • fa3589e5c5 Update our error messages to point to getsatisfaction.com/gsa. hanna 2010-01-27 19:16:28 +0000
  • 3399ad9691 Incremental update 2 -- refined allele and VariantContext classes; support for AttributedObject class; extensive testing for Allele class, and partial for VariantContext. Now possible to easily convert dbSNP to VariantContext. depristo 2010-01-27 17:19:37 +0000
  • 3edcefb7fb add _gI and _gD to the indel probe names according to the spec (in the hope that wiki is not obsolete); added optional cmd line param -project_id to prefix all probe names with. asivache 2010-01-27 17:06:49 +0000
  • ed9b7edee3 Changed " to ' to stop the chartl 2010-01-27 15:23:55 +0000
  • 40c242d2b8 Fix for overflow issues depristo 2010-01-27 13:37:16 +0000
  • 8453676b71 added a method to AlignmentContext called hasExceededMaxPileup, which you can use to determine if the current site exceeded the maximum pileup size (reads were dropped). Added this as a check to unified genotyper according to Eric's instructions, and added the plumbing to the engine. aaron 2010-01-27 05:17:01 +0000
  • 4bcdab580c --output_dir has been changed to --output_prefix to give the user more control over the names of the resulting mass of files in AnalyzeAnnotations. The fontsize of the axes is increased. Cumulative filtering plots are removed since the binned filtering plots are much more useful. rpoplin 2010-01-27 04:50:54 +0000
  • df112e64b8 Minor tweaks chartl 2010-01-27 04:17:47 +0000
  • bbddeb693f Updated the package to use RealignerTargetCreator. Also, removed IndelGenotyper since it's still experimental. ebanks 2010-01-27 03:45:15 +0000
  • 476d6f3076 RealignerTargetCreator is officially live ebanks 2010-01-27 03:41:52 +0000
  • c1f154c31f Fully qualify the target for the output file so that external build processes (aka firehose) won't aren't impacted when building the GATK from outside the base directory. hanna 2010-01-26 23:06:30 +0000
  • 169032aa78 Add new 'release' task to push releases into a shared directory and the website. hanna 2010-01-26 22:50:20 +0000
  • 1f64c5d41a Do not slurp the whole set of snp mask sites into memory (gets pretty heavy on full dbSNP!); instantiate a privare ROD iterator instead and drag it across the sites we are designing probes for. asivache 2010-01-26 22:39:46 +0000
  • 47440bc029 - Removed max_coverage argument from UG; Aaron will set it up so that we don't call when the GATK had to drop reads. - Reimplemented optimization in UG to not call when there are no non-ref bases. - Compute reference confidence accurately in UG for ref calls. ebanks 2010-01-26 21:56:33 +0000
  • 2c8d7b0c44 Forgot the onTraversalDone. That was dumb. chartl 2010-01-26 21:02:46 +0000
  • 04e1832968 Added - AlleleBalanceHistogramWalker -- hopefully this'll be able to tell us very clearly whether bad genotype concordance is a result of systematic contamination (consistent wonky allele balances) chartl 2010-01-26 20:57:12 +0000
  • a1054efe8a Default platform and default read group are no longer set to values by default. The recalibrator throws an exception if needed values are empty in the bam file and the args weren't set by the user. This is done to make it more obvious to the user when the bam file is malformed. Similarly, the recalibrator now refuses to recalibrate any solid reads in which it can't find the color space information with an exception message explaining this. The recalibrator no longer maintains its own version number and instead uses the new global GATK version number. rpoplin 2010-01-26 18:47:40 +0000
  • 31d252c0a1 Now using the new VariantEval package. kshakir 2010-01-26 16:32:44 +0000
  • 0345d9f6a5 Updating the recalibrator to use non-depricated getPileup() method. Adding documentation to AnalyzeAnnotations so that the walker isn't marked as unclean at compile time. rpoplin 2010-01-26 14:15:09 +0000
  • c231547204 Refactoring and migration of new allele/variantcontext/genotype code into oneoffprojects. NOT FOR USE. PlinkRod commented out due to dependence on this new, rapidly changing interface. depristo 2010-01-26 13:53:29 +0000
  • 3380b6ebe8 Clean up packaging in preparation for actually releasing. hanna 2010-01-26 03:12:49 +0000
  • 2e57bc7879 added a better message for the SO flag error in MergingSAMIterator2 aaron 2010-01-25 22:57:18 +0000
  • 24d4082925 AnalyzeAnnotations can now process only variants that are found in samples that match the -sampleName argument. X-axis of plots no longer use annoying scientific notation. rpoplin 2010-01-25 20:52:11 +0000
  • 022601b1a5 Warnings for walkers w/o Javadoc. hanna 2010-01-25 20:34:50 +0000
  • 894a2b511b Fixing no platform warning message. rpoplin 2010-01-25 19:46:50 +0000
  • 2b51cf18f0 AnalyzeAnnotations now outputs plots with log x-axis in addition to standard x-axis so things like DP and MQ0 are easier to see. AnalyzeAnnotations now skips over all annotations that aren't floating point values. Recalibrator now warns users if PL tags are missing and so therefore it is reverting to illumina. rpoplin 2010-01-25 19:39:18 +0000
  • 6cf413e630 Bug: ExpandedSAMRecord did not treat hard-clipped bases ('H') correctly. Fixed. asivache 2010-01-25 19:23:44 +0000
  • 4990139b60 A collection of python objects that are useful for VCF validation. Use 'em or don't. chartl 2010-01-25 18:44:10 +0000
  • dc170caafc Now, if a dbsnp rod is passed to either the UnifiedGenotyper or VariantAnnotator, a DB=0/1 annotation is added (in addition to filling in the ID field); this is in line with 1KG project calls. If no dbsnp rod is used, the annotation is not added (as opposed to setting every entry to DB=0). ebanks 2010-01-25 17:27:12 +0000
  • 5d2f8aaa54 Updating recalibrator version number after the several emergency changes last week. rpoplin 2010-01-25 14:35:47 +0000
  • 588417e17d Don't reference that optimiation library I'm not using anyway. jmaguire 2010-01-24 20:30:50 +0000
  • d3e3c1c2e0 don't require that optmization lib that I'm not using yet... (doh) jmaguire 2010-01-24 20:28:21 +0000
  • 1d6d2b26f7 tools for optimizing calls. jmaguire 2010-01-24 20:16:55 +0000
  • 877957761f lots of new stuff, some generally useful, some one-off. jmaguire 2010-01-24 19:50:48 +0000
  • 78890c0bee First version of walker that combines the functionality of IndelIntervalWalker, MismatchIntervalWalker, SNPClusterWalker, and IntervalMergerWalker - plus it allows the user to input rods containing known indels (e.g. dbSNP or 1KG calls) for automatic cleaning. Basically, all pre-processing steps for cleaning are now done in a single pass. More testing needed. ebanks 2010-01-24 05:32:38 +0000
  • d6b9b788a8 Renamed -- PlinkRodWithGenomeLoc --> PlinkRod chartl 2010-01-23 18:19:28 +0000
  • ac983e7a0b Ran the rod on a binary plink file with indels and it just worked. Love it when that happens! Unit test to ensure this behaviour is maintained. chartl 2010-01-23 18:13:05 +0000
  • ae22d35212 PlinkRod now correctly parses binary files without indels; unit test added for this behavior. chartl 2010-01-23 17:34:06 +0000
  • 94dc09c865 PlinkRod now successfully instantiates on the binary ped file trio (.bim, .bam, .fam) for non-indel files. chartl 2010-01-23 16:13:24 +0000
  • 01db93299c PlinkRodWithGenomeLoc now properly handels indels. chartl 2010-01-23 07:34:52 +0000
  • 42fb85e7f3 PlinkRodWithGenomeLoc now properly parses text plink files. Unit test added to test this functionality. Indels and binary files to come. chartl 2010-01-23 06:19:26 +0000
  • 648a36d08e Temporary solution: add the commons logging implementation to the VariantFiltration package. Downstream solution is described in GSA-262. hanna 2010-01-23 01:46:18 +0000
  • c871a0f221 UG map() now returns a VariantCallContext object. Also has a field for confidentlyCalledBases. UG reduce() emits statistics on the confident called % of bases depristo 2010-01-22 23:06:43 +0000
  • fbf82526cb Minor renamign changes. chartl 2010-01-22 23:06:32 +0000
  • fd223e955c Reverting the previous solid change. We now refuse to recalibrate if the solid read doesn't contain proper color space information. The exception message has been updated to say this. Also, Tile has been downgraded to an ExperimentalCovariate due to performance issues. rpoplin 2010-01-22 20:55:28 +0000
  • 7732f98e56 Fix for Solid reads that have '.' in their color space field. The recalibrator will just set them to be illumina reads and won't apply color space correction. rpoplin 2010-01-22 20:09:16 +0000
  • 2ea768d902 ant clean is your friend....fixed test code dependent on an interface change. aaron 2010-01-22 20:07:46 +0000
  • a11503819a AnalyzeAnnotations now breaks out its TiTv plots into novel SNPs, dbSNP sites, and combined. rpoplin 2010-01-22 19:00:23 +0000
  • e00cb688ac Cleanup of GATK-GSA-Pipeline to support the new naming. hanna 2010-01-21 22:17:35 +0000
  • cc3b818268 cleanup of the pile-up limit exceeded warning, and a little code cleanup aaron 2010-01-21 22:17:24 +0000
  • de21943acd A few more bug fixes based on extended testing. Sorry, Eric. hanna 2010-01-21 22:12:16 +0000