2183f10a1dScript for validating and converting text files into the tabular format required for GenomicAnnotator -B inputs
weisburd
2010-04-13 13:35:10 +0000
e413882302Generalizing the SequenomValidationConverter to be able to take in any arbitrary rod type (provided it can be converted to VariantContext).
ebanks
2010-04-12 20:42:18 +0000
4bb8984f80Updating picard -- switching to Alec's more robust fix for gzip decompression issue and updating serialization components.
hanna
2010-04-12 19:56:45 +0000
14b8101d45Error message fail. Failed to supply one of the valid interval file types.
hanna
2010-04-12 01:19:01 +0000
60d54e69f3Hackish fix to present a better error message if the file does not have the proper extension. Will work with Brett to come up with a better solution.
hanna
2010-04-12 01:11:27 +0000
d06c7835d8Adding performance tests for the indel realigner; should take ~3 hours.
ebanks
2010-04-11 04:45:22 +0000
3434a61146Don't trigger when ref=N (which can happen when a dbsnp track is provided)
ebanks
2010-04-11 02:59:11 +0000
961ca05abcRemoved outdated Sequenom rod and renamed HapMapGenotypeROD to HapMapROD.
ebanks
2010-04-11 01:43:07 +0000
fa01876255UnifiedGenotyper performance tests (WG, WEx); currently takes just over an hour.
ebanks
2010-04-09 19:42:29 +0000
0cc6d0fbbbOne more quick memory improvement: reuse Alleles in a given context instead of creating new ones for each sample (duh).
ebanks
2010-04-09 18:48:36 +0000
c2a37e4b5cVariant Quality Score modules in VariantEval2 no longer create huge lists which hold all of the quality scores encountered and instead cast the quality score to an integer and use hash tables. Bug fix for files in which all the quality scores are set to -1.
rpoplin
2010-04-09 18:36:06 +0000
71f38a9199Adding performance tests for the recalibrator (Whole Genome and Whole Exome tests). Should take ~3 hours to run.
ebanks
2010-04-09 18:30:59 +0000
e73e6a4fb0Significant memory improvements to plink code
ebanks
2010-04-09 16:12:38 +0000
75c1987a18fixed the build
aaron
2010-04-09 13:03:31 +0000
f1b1e70612Bug fix for multisample calls in ApplyVariantClusterWalker
rpoplin
2010-04-09 12:01:15 +0000
3f2455e346Better error message as suggested by James P
ebanks
2010-04-09 05:52:53 +0000
fba48b515aHeads up everyone: For consistency, these tools should be writing to the walker's output stream and no longer use the -vcf argument.
ebanks
2010-04-09 05:37:25 +0000
e286623f6fUse byte[] instead of String in an attempt to cut down on memory usage
ebanks
2010-04-09 05:32:54 +0000
7025f5b51dAdded an auxiliary table to DepthOfCoverage, which is the cumulative equivalent of the locus table (got tired of doing the calculation by hand). Also took care of a trailing tab in the per-locus output table.
chartl
2010-04-08 19:37:17 +0000
fab31e1d53Check in so I don't lose this code -- spawning of jobs by genes
chartl
2010-04-08 16:18:40 +0000
9f6377f7fbadded a performance test build option (for the upcoming performance test suite), and added a sample performance test for VariantEval.
aaron
2010-04-08 15:37:15 +0000
4014a8a674A long overdue correction; all unit tests now end in 'UnitTest'. This was something we wanted to do for a while, and now with the performance tests coming, it was a good time to clean-up. Please label any new test appropriately: *UnitTest and *IntegrationTest are the two valid file name patterns for tests.
aaron
2010-04-08 06:14:15 +0000
e148a3ac61added the ability to create interval lists directly from a ROD, using the command line arg '-BTI' (long name '--rodToIntervalTrackName'). The parameter to this arg is the name of the ROD track, which must be a track name specified in the -B option.
aaron
2010-04-08 05:14:41 +0000
20cc2a85a4removed the hashmap from Genotype Concordance, moved it into a table
aaron
2010-04-07 21:24:48 +0000
e55f27b3b1forgot a file
aaron
2010-04-07 20:51:13 +0000
9ca8e345fcby-by old junk.
aaron
2010-04-07 20:41:48 +0000
8fd59c8823Modified the report system based on Ryan's feedback: tables are now created independently to avoid the permutation problem when they were all compressed in rows, and removed our dependency on FreeMarker. The Grep format stays the same.
aaron
2010-04-07 20:39:55 +0000
918b746798More detailed validation output. Fixes for genotyping overflow -- these are temporary and need to be properly resolved
depristo
2010-04-07 16:38:28 +0000
e7dad728dfTrivial output changes for consistency
ebanks
2010-04-07 14:47:43 +0000
058e7d3d12Bug fix for Gregory
depristo
2010-04-07 00:21:35 +0000
7b44e6bd55ApplyVariantClusters now outputs interesting threshold points based on hitting the target novel TiTv
rpoplin
2010-04-06 19:47:29 +0000
60c227d67fAdded new VE2 module to create a plot of titv ratio by variant quality score
rpoplin
2010-04-06 15:19:27 +0000
3530ef5a41Explicit type cast fixed in order to work with new ROD implementation
asivache
2010-04-06 15:02:56 +0000
2d002c56c3Added histogram of variant quality scores broken out by true positive and false positive calls to the GenotypeConcordance module of VariantEval2
rpoplin
2010-04-06 13:48:31 +0000
12e4f88ca7a little bit more clean-up
aaron
2010-04-05 20:49:06 +0000
df7e7921ceremoving some unused code.
aaron
2010-04-05 19:30:08 +0000
56eb15f91fError checking for bad input (thanks, Aaron).
ebanks
2010-04-04 03:17:01 +0000
27fb6f7594Make sure to convert non-integer chromosomes (M,X,Y) back from their corresponding integer representations (0,23,24) when writing in .bed format
chartl
2010-04-03 20:01:21 +0000
705b28e90dFirst attempt at implement record filtering based on special 'hap_ref', 'hap_alt' columns in the input files
weisburd
2010-04-02 21:52:26 +0000
8017fb123fchanged the depth of coverage walkers class name, and added a dependency in the packaging system so that RODs will all get imported.
aaron
2010-04-02 20:55:19 +0000
5a3622b051Get rid of CoverageStatistics from packages so ant package doesn't go looking for it.
chartl
2010-04-02 18:04:33 +0000
6b7b07f178First checkin of GenomicAnnotator which annotates an input VCF file by pulling data in a generic way from an arbitrary set of TabularRODs.
weisburd
2010-04-02 17:49:42 +0000
699a0ea9d1adding SequenomValidationConvertor and VariantsToVCF
ebanks
2010-04-02 17:27:42 +0000
687fd477ffJust some code I want to freeze. If you ever need to estimate the % of bases covered by exon, given an interval list, give it to getTargetedGenes. Not the best name for this function, but I don't expect anyone to use it but me.
chartl
2010-04-01 20:21:50 +0000
d7880ef7adForgot to uncomment the AlignerIntegrationTest before committing. And yes, matt, commenting it out is, in fact, easier than just setting my classpath.
chartl
2010-04-01 17:17:16 +0000
f7d1b8f5deCoverageStatistics has now replaced DepthOfCoverage -- old DoC is in the archive.
chartl
2010-04-01 16:27:23 +0000
585cc880a2changed jexl expressions to jexl names in the VariantEval2 output, fixed integration test, and fixed a problem where a line was getting dropped in CSV output
aaron
2010-04-01 16:23:14 +0000
d00bde22dbReverting one of Brett's changes that should not have been committed. Will address with Brett separately.
hanna
2010-04-01 16:10:46 +0000
b4f6f54502Reorganizing the way interval arguments are processed Most of the changes occur in GenomeAnalysisEngine.java and GenomeLocParser.java: -- parseIntervalRegion and parseGenomeLocs combined into parseIntervalArguments -- initializeIntervals modified -- some helper functions deprecated for cleanliness Includes new set of unit tests, GenomeAnalysisEngineTest.java
bthomas
2010-04-01 12:47:48 +0000
c3c6e632d1support for two new VCF header info field value-types, Flag (for fields that are just boolean truths), and Character (for single charatcer info fields).
aaron
2010-04-01 03:11:32 +0000
3d3d19a6a7the last-mile commit for Tribble integration. The system is now ready for Tribble to be turned on, as soon as we've removed any dependencies in the ROD code on interfaces that aren't in the Tribble library (i.e. the Variation or Genotype interface on RODs). All of the walkers should be up to date.
aaron
2010-03-31 22:39:56 +0000
4fcee248f9For Kristian: functions which, given a read, can uniquely identify the BAM file storing that read. Introducing this into the pile of code which peeks under the covers of the SAMDataSource in the hopes that this function can help to replace the others and provide a single path for crosstalk.
hanna
2010-03-31 20:46:44 +0000
ac9c335cd2This is a python job dispatcher I've been using, which builds on Mark's FarmJob utility, and an example script of how I'm using it. Basically I wrote it to smartly break up analysis over an interval list, givin a maximum number of bases per job, a list of available queues, and a limit on each queue. It handles going over these limits in three ways:
chartl
2010-03-31 19:53:13 +0000
d58fe70708Correctly ignore filtered calls and indel calls in the truth sets
rpoplin
2010-03-31 14:33:01 +0000
b60197ae10Another round of cleanup and simplification in Picard -- Picard's unit tests are now passing for my branch.
hanna
2010-03-31 01:02:59 +0000
40f8e7644cBetter, multi-haplotype aware haplotype scores. Looking very good now, seems to be vastly better at dealing with incorrect calls in deep and low pass data. Almost ready for use
depristo
2010-03-30 23:57:36 +0000
f992f51a3bDeleting incorrect sampling genotype likelihoods from the codebase
depristo
2010-03-30 23:56:35 +0000
b9d3fc3fbbNow checks if the i-th element of the FiltrationContext[] is null before trying to access it. This seems to happen occassionally at the very end of a VCF file... the array will be 6 elements long, but the last element will actually be null.
kiran
2010-03-30 22:40:17 +0000
400684542cRevisions to take into account finalization of Picard patch: naming changes, better definition of public interfaces. This won't be the last Picard patch, but it should be the last big one.
hanna
2010-03-30 19:28:14 +0000
b00d2bf2bcfixing an annotation that was breaking the error log output system.
aaron
2010-03-30 15:34:04 +0000
a6e8687d71implementing a clean way to import the template files into the GATK jar (they should not always get bundled). All further resources should be added to the gatk.resources path id in the build script.
aaron
2010-03-30 04:20:19 +0000
babb9fb825snp cluster filter should ignore ref calls when determining the clusters
ebanks
2010-03-29 17:57:33 +0000
24461a2503Let's *not* import classes that no longer exist. How my own ant test compiled is beyond me.
chartl
2010-03-29 13:59:01 +0000
dc802aa26fMoved CoverageStatistics to core. This will be (soon) renamed DepthOfCoverage; so please use CoverageStatistics
chartl
2010-03-29 13:32:00 +0000
8ea98faf47Deleting the pooled calcluation model -- no longer supported.
depristo
2010-03-29 11:44:27 +0000
85037ab13fFix for Kiran's sharding issue (Invalid GZIP header). General cleanup of Picard patch, including move of some of the Picard private classes we use to Picard public.
hanna
2010-03-29 03:21:27 +0000
b8ab74a6dcMinor useful changes to BaseUtils and MathUtils to support a new haplotype score annotation that determines to the two most likely haplotypes over an interval and scores variants by their consistency with a diploid model. Appears to be useful.
depristo
2010-03-28 21:45:22 +0000
e9e53f68abFilter lists can now end with .list or .txt.
kshakir
2010-03-27 17:41:24 +0000
074ec77dccFirst go of the new output system for VE2. There are three different report types supported right now (Table, Grep, CSV), which can be specified with the reportType command line option in VE2.
aaron
2010-03-27 03:59:32 +0000
85f4f66180Updated to use VariantContext. Output has been reformatted: variant and genotype concordance are emitted for every coverage level per variant. If the requested sampling level is higher than what's available, the maximum available coverage at that locus is used. This makes it much easier to make plots indicating the percentage of comparison callset recovered at a certain sampling depth.
kiran
2010-03-26 21:02:43 +0000
391e5843e4If the annotation engine has not been supplied, don't try to annotate anything.
kiran
2010-03-26 20:52:21 +0000
8048b709a0Selects a single sample on which to operate.
kiran
2010-03-26 20:50:58 +0000
20e3ba15caAdded an optional argument -rgbl --read_group_black_list to filter read groups.
kshakir
2010-03-26 19:38:57 +0000
73a14a985bMoving VariantsToVCF to core.
ebanks
2010-03-26 18:55:12 +0000
14bf6923a8HapMap-to-VCF now works fine within Variants-to-VCF. Added integration test for it and removed old code.
ebanks
2010-03-26 18:34:59 +0000
78af6d5a40New sharding system is going live again for on-the-fly merging.
hanna
2010-03-25 18:39:04 +0000
46c14ec63fNew, much less memory intensive implementation of BAM file sharding. Streams indices together with the expectation that bins will be present in the bin sparse array, which avoids the problem of having to hold the sparse bin array stored in every BAM file index in memory at the same time.
hanna
2010-03-25 17:41:22 +0000
4398a8b370Updated. Now uses VariantContext and is truly "variants" to vcf (i.e. not just GELI to vcf).
ebanks
2010-03-25 04:53:31 +0000
2373a4618fbug caused by a misprint: context != contexts
ebanks
2010-03-25 03:08:24 +0000
3176715c741. Alignability mask returns null when not available. 2. --list now prints out the available classes/groups too.
ebanks
2010-03-24 20:49:07 +0000
06a212e612Adding VariantConcordanceROCCurveWalker to create ROC curves comparing concordance between optimized call sets and validation truth sets in VCF format in order to evaluate performance of variant optimizer independently of achieving a particular novel ti/tv ratio. Added option to ignore only the specified filters in the input call sets via --ignore_filter <String>. Added option to provide a prior estimate of error for known snps via --known_prior <qual>. The het and hom calls are clustered independently. Infrastructure in place to use titv of known snps to inform p(true) of novel snps. Tweaked protection against overfitting based on suggestions from several people. Minor edits to AnalyzeAnnotations.
rpoplin
2010-03-24 19:43:10 +0000
47e30aba92Rods for reads hooked up into the cleaner
ebanks
2010-03-24 18:17:56 +0000
5079f35e40better method names for read based reference ordered data access.
aaron
2010-03-24 16:13:31 +0000
49117819f5For the cleaner to clean, it must beat the entropy produced by the aligner (and not just the raw reads).
ebanks
2010-03-24 15:21:58 +0000
60dfba997badded some sample annotations to VariantEval2 analysis modules, and some changes to the report system.
aaron
2010-03-24 05:40:10 +0000
1f451e17e5Changing preloaded index to only "preload" reference sequences on demand. Results in drastic lowering of startup cost when multiple BAM files are merged.
hanna
2010-03-23 22:02:28 +0000
884a577013Phase 2 of Picard patch refactoring: kill off SAMFileReader2/BAMFileReader2, merging the changes back into the base classes.
hanna
2010-03-23 16:48:11 +0000
7462a0b2d1cleaned-up of VariantContextAdapter tests, fixed the double comparisons in equals() in RodGeliText (nice MathUtils.compareDoubles Kiran)
aaron
2010-03-23 15:18:30 +0000
a69b8555ddGeli to variant context.
aaron
2010-03-23 06:45:29 +0000
eafdd047f7GLF to variant context. Added some methods in GLF to aid testing; and added a test that reads GLF, converts to VC, writes GLF and reads back to compare.
aaron
2010-03-23 03:43:25 +0000
3767adb0bbProcessing intervals as they stream in means much lower memory usage and quicker runtime. Making change as minimal as possible to avoid conflicts with BT's incoming patch.
hanna
2010-03-22 22:04:45 +0000
0097106938VariantFiltration can now filter specific samples. This is *NOT* an ideal implementation. One day when we have lots of free time (or a greater desire), we will implement this correctly and sophisticatedly using all the power of JEXL. For now, though, this will have to do. Docs coming tonight.
ebanks
2010-03-22 20:45:11 +0000
543aefc3d7Fixing the bug introduced with the earlier commit. When trimming locus to the current bases, we need to take into account expanded boundaries (for windowed reference traversals)
asivache
2010-03-22 19:20:34 +0000
ee1dc6092fTest updated. Now we do not throw an exception when locus interval is out of bounds, we just return silently a reference context trimmed to the current shard boundaries. New test checks for trimming.
asivache
2010-03-22 17:37:52 +0000
d2944461efWe also have to allow the window to be (partially) outside the bounds and trimming to the contig size is not enough (thanks to shards). Now we trim to the current bounds too (i.e. if the interval is not completely within current bounds, we create reference context that contains only bases from the overlap between the interval and the bounds).
asivache
2010-03-22 17:36:29 +0000
weisburd
ebanks
hanna
rpoplin
aaron
chartl
depristo
asivache
bthomas
kiran
kshakir