cc2bf549c8Removing my unnecessary optimization. 10 lines later in the code the same optimization was applied. A monumental waste of time.
depristo
2010-05-28 14:10:48 +0000
a4d834cc01fixing the test I broke
aaron
2010-05-28 02:06:20 +0000
6485e8383dTrivial change to retrigger broken build that really isn't broken
depristo
2010-05-27 23:33:46 +0000
f2e7582cfcReorganization of SW code for clarity. Totally failure at raw optimization. Discovered that ~50% of reads being cleaned were perfect reference matches. New code comes with flag to look at NM field and not clean perfect matches. Can we turned off with command line option (needed for 1KG bams with bad NM fields). Going to rerun cleaning jobs due to accidentally rebuilding of stable codebase and loss of 2 days of runtime.
depristo
2010-05-27 23:16:00 +0000
e1b0aefb29fix for parallelism bug
aaron
2010-05-27 22:16:14 +0000
cded9ec985adding a command line option, -etd (enable threaded debugging), that uses a custom thread pool class to catch exceptions thrown inside of a thread.
aaron
2010-05-27 21:57:56 +0000
0e3c76ab84GSA-333 Meant to use getAbsolutePath and not deref symlinks.
kshakir
2010-05-27 21:23:02 +0000
e2674671e7The liftover code needs to *hard filter* records whose reference changes (since they no longer adhere to the VCF spec as they don't match the new reference - and can't be converted to VariantContext).
ebanks
2010-05-27 19:22:47 +0000
ff4a0764dfRead error rate is now parallelizable
chartl
2010-05-27 19:00:09 +0000
99aae4ac0bended not done for LSF queue submission.
depristo
2010-05-26 22:16:48 +0000
dfc36c1e95Restructuring of the mandatory read filters for traversals. Now everything uses ReadFilters, even for the required filters like being mapped for LocusWalkers. Statistics now tracked for each read filter used during the traversal and info emitted in INFO at the end.
depristo
2010-05-26 22:12:25 +0000
3873dccb35First fully functional (though preliminary) version of walker that takes an input VCF and outputs a Beagle .bgl file that can be used for missing genotype calls/haplotype imputation. For now, only supported input format is likelihood format for unrelated individuals.
delangel
2010-05-26 21:03:23 +0000
f9efc1248cVariantEvalWalker now takes indels if you throw the -dels flag. IndelLengthHistogram appears to be working properly, it is turned off by default (as it is experimental) but you can turn it on in your own repository.
chartl
2010-05-26 20:03:14 +0000
a0bf1753f9A script to sanity check that a release is runnable. Not fancy - just pulls out the performance tests in 20sec chunks. Only a temporary solution.
ebanks
2010-05-26 17:45:42 +0000
058441fa39Trivial renaming of test
ebanks
2010-05-26 16:56:42 +0000
0265199ce4First pass at an IndelLengthHistogram module for variant annotator. Off by default. Will be tested shortly (have to commit, so I can check out in another directory, so that compiling won't kill all my jobs running on LSF)
chartl
2010-05-26 15:04:39 +0000
a2fab07258fixed the build problem: there were two copies of the AnnotatorInputTable Codec and Feature in two different spots.
aaron
2010-05-26 14:47:15 +0000
5928047d8bOptimization of reference window calculation to us bytes not char and no uppercasing since reference and read bases are always uppercase now. Should remove some ~5% of runtime of UG.
depristo
2010-05-26 14:10:26 +0000
88a06ad81fChanges to Depth of Coverage:
chartl
2010-05-26 03:39:22 +0000
0607f76a15commenting out this test until I can figure out what the hell is going on with the codecs.
ebanks
2010-05-26 01:12:10 +0000
ed4d8ddd05Refactoring and sugar to give lower level access to the job graph. Will add more sugar / glue depending on how much of a graph the python generator outputs.
kshakir
2010-05-25 22:52:29 +0000
062b316881Better Exception message when can't find annotation value in variant recalibrator.
rpoplin
2010-05-25 21:15:50 +0000
bf530d23deVariant Recalibrator now makes use of a prior on known/novel status as well as on allele frequency spectrum. The VariantOptimizer walker now clusters with all variants but gives more weight to knowns / hapmap / 1KG / MQ1 sites. The weights are all optional command line arguments. We no longer assign default values to annotations that are malformed. The walkers will crash with exception so as to not cover up potential issues. We only produce titv-less clusters now, and so the titv argument in VO was removed and the WithoutTiTv string that gets added to the cluster file is removed. The wiki is updated to show new example commands.
rpoplin
2010-05-25 21:08:31 +0000
ae6c014884Fixed UG parallelization bug. Better integration test to catch this in the future.
ebanks
2010-05-25 21:03:45 +0000
434e920da9Oops, forgot to update integration tests
ebanks
2010-05-25 20:37:45 +0000
772f558ae0Massive change to the indel realigner code. We now properly deal with soft-clipped reads. Also, improved left-alignment code. Small change for Ryan to get hard-clipped reads working for the recalibrator.
ebanks
2010-05-25 20:04:33 +0000
f3e2aae570add experimental support for tabix files (for any of our Tribble rod types), as long as they end in .gz and can be read by the tabix reader.
aaron
2010-05-25 04:44:46 +0000
f7c9f131earevisioning tribble to version 85, which includes tabix and bug fixes
aaron
2010-05-25 04:28:32 +0000
8db7c97c4dMoved AnnotatorInputTableFeature and Codec to org.broadinstitute.sting.gatk.refdata.features.annotator
weisburd
2010-05-24 14:38:54 +0000
4aa749c709Moved AnnotatorInputTableFeature and Codec to org.broadinstitute.sting.gatk.refdata.features.annotator
weisburd
2010-05-24 14:38:07 +0000
aca3bcb193Moved AnnotatorInputTableFeature and Codec to org.broadinstitute.sting.gatk.refdata.features.annotator
weisburd
2010-05-24 14:37:17 +0000
64ed770250Moved AnnotatorInputTableFeature and Codec to org.broadinstitute.sting.gatk.refdata.features.annotator
weisburd
2010-05-24 14:36:28 +0000
42fa12d9e4without the *, the package system only took directly referenced files (which pulled in VCF and BED) but skipped dbsnp. Epic star fail...
aaron
2010-05-24 05:19:57 +0000
ee3f2eb1d0Don't output traversal reduce result in the logger. In many cases, the reduce result is tangential to the product of the analysis and having the logger always emit it can confuse the output (such as in the new reduceByInterval DepthOfCoverage walker). If users want to emit it, they can choose not override onTraversalDone, or override onTraversalDone and write results to the output stream / logger / whatever their choice.
hanna
2010-05-23 22:41:43 +0000
a40e64e47bA downsampling validator. Compares the generated pileup passed in from the alignment context to the reads, passed in as a Tribble SAM text feature. If the generated pileup contains a valid set of reads according to the downsampling rules, the test passes.
hanna
2010-05-23 21:49:54 +0000
e9ee55d7ddA cleaned up functioning early, early access version of Queue for others to play with and provide feedback about next steps. Current version only has syntatic sugar for accessing the graph via rules ex. "bam" -> "bam.bai", "samtools index ${bam}" and DOES NOT have sugar for constructing your own graph. Usage info on the internal wiki at https://iwww.broadinstitute.org/gsa/wiki/index.php/Queue
kshakir
2010-05-23 20:21:09 +0000
a280a0ff0da) Made HaplotypeScore default annotation. This changed several integration tests, whose MD5 is now updated. b) Disabled BaseQualRankSumTest, the returned p-values differ wildly from Matlab/R-provided ones, cause TBD.
delangel
2010-05-21 22:25:17 +0000
f32a32269cminor change for eric
depristo
2010-05-21 20:13:36 +0000
b10950c691Simple performance optimization -- cache the number of reads in the locus hanger.
hanna
2010-05-21 19:26:16 +0000
355396109bBug fix to avoid build failure (class changed under me??)
delangel
2010-05-21 18:48:56 +0000
1753d07b02Added AnnotationByAlleleFrequencyWalker - walker takes an input vcf, a reference vcf and a list of annotations (with the -A argument). For each site present in both VCF's, it outputs the given annotations into the screen as well as allele frequency. Since HapMap vcf reference doesn't include AF in annotations, it computes it from Chromosome, Het and HomVar counts.
delangel
2010-05-21 18:31:34 +0000
745d7c582fadded integration test for intervals with no coverage due to filtering
chartl
2010-05-21 16:52:42 +0000
7fb3f2d3ebAnnotator now buffers indel calls (prevents double-output from double-calls to map)
chartl
2010-05-21 16:34:34 +0000
4e834b5e35VFW now uses a ref window and thus is compatible with indels.
chartl
2010-05-21 15:59:42 +0000
7910a6a7cbadding all of Tribble, not just select codecs, to the package
aaron
2010-05-21 15:48:50 +0000
d1098fa77bRemoving unnecessary dependences that were causing problems for Sendu
depristo
2010-05-21 13:07:41 +0000
88cb93cc3cChanges to Depth of Coverage (added maximum base and mapping quality flags; with new integration tests -- because they use b36, and the other test uses hg18, it's in a different class (integration test system can't change refs on the fly). Initial change to VariantAnnotator to allow it to see extended event pilups; you currently have to throw the -dels flag; and it's specified as "very experimental". Yet,all the integration tests pass.
chartl
2010-05-21 13:02:01 +0000
6faf101c6cMinor improvements to Callable Loci for public consumption
depristo
2010-05-21 12:50:11 +0000
388dd8d64dFixing bugs in downsampler introduced when I added Ryan's dup eliminator.
hanna
2010-05-21 02:53:12 +0000
a10fca0d5cGenotyper now is using bytes not chars. Passes all tests.
depristo
2010-05-20 21:02:44 +0000
7389077b3bA few misc usability fixes: - Clarify the message emitted when -XL is supplied so I don't spend another half day chasing a bug that doesn't exist. - Crash with a helpful message when running -nt with non-TreeReducible walkers. - Crash with a helpful message when running -nt with reduceByInterval walkers.
hanna
2010-05-20 19:02:02 +0000
b543dd4ac4more aggressive checks for the locking, and some more documentation
aaron
2010-05-20 16:16:36 +0000
fc57bccde8trivial removal of directory
depristo
2010-05-20 14:25:32 +0000
727822adb4BaseUtils has more clear distinction between byte and char routines. All char routines are @Depreciated now. Please use bytes. Better organization of reverse(), now in Utils not BaseUtils.
depristo
2010-05-20 14:05:13 +0000
6ce3835622Removing unused methods in QualityUtils; ReferenceContext now converting all bases to upper case, but can be disabled with static boolean
depristo
2010-05-20 12:38:06 +0000
5abac5c057A few more char -> byte cleanups
depristo
2010-05-20 00:02:06 +0000
8a725b6c93Restructuring of ReferenceContext and ReadWalkers to accept a ReferenceContext. Now ReferenceContext is byte[] backed not char[]. Please no more chars for the reference. All of the tests pass now. Coming check-ins are going to clean up the char / byte problems in the GATK
depristo
2010-05-19 23:27:55 +0000
02cc1afdc8remove RodBed and all it's dependencies.
aaron
2010-05-19 19:12:30 +0000
ffb1b46166Added a GCCalculatorWalker for a oneoff analysis for Mark Daly (GC content of agilent 1.1 targets)
chartl
2010-05-19 18:49:51 +0000
0036df7b03adding a convenience method for getting at the RODs that overlap a specific locaiton as GATKFeatures.
aaron
2010-05-19 17:40:20 +0000
ca386439beonly emit a warning if the tribble index is out of date, don't remove and replace it for them. Added a test case where the log4j appender checks the logging messages for the appropriate output.
aaron
2010-05-19 15:12:48 +0000
017ab6b690Experimental versions of downsampler and Ryan's deduper are now available either as walker attributes or from the command-line. Not ready yet! Downsampling/deduping works in a general sense, but this approach has not been completely optimized or validated. Use with caution.
hanna
2010-05-19 05:40:05 +0000
46ba88018dUpdated to the new readHeader(..) api
weisburd
2010-05-19 04:06:34 +0000
984c51efd3Updated to use Tribble-based GATKFeature instead of TabularROD
weisburd
2010-05-19 03:42:12 +0000
42ee16f256Updated to use Tribble-based GATKFeature instead of TabularROD
weisburd
2010-05-19 03:41:37 +0000
d8469e2fbaUpdated to use Tribble-based GATKFeature instead of TabularROD
weisburd
2010-05-19 03:40:47 +0000
d65b2d32d1Removed AnnotatorROD which has been ported to Tribble
weisburd
2010-05-19 03:39:34 +0000
b82116f488Removed AnnotatorROD which has been ported to Tribble
weisburd
2010-05-19 03:39:20 +0000
6b96f025f5Tribble integration for indexing the AnnotatorInputTable format
weisburd
2010-05-19 03:37:54 +0000
2f3933148dAdded fast split(str, delimiter) methodf
weisburd
2010-05-19 03:37:26 +0000
aedb9f6734Bring SAMPileupCodec into compliance with new interface.
hanna
2010-05-19 01:23:29 +0000
7cfb9ff3dcupdates for Tribble 82, fixes for Ryans case where multiple processes would attempt to read/write to the same index, and a couple other Tribble-centric bug fixes.
aaron
2010-05-18 19:34:45 +0000
635f61c22dClone the other guy too
chartl
2010-05-18 18:56:01 +0000
9e15299475Misc cleanup in variant recalibrator.
rpoplin
2010-05-18 17:37:01 +0000
eb200e4cceHrumph. Don't just add pointers to the same objects, actually clone the underlying arrays.
chartl
2010-05-18 17:13:44 +0000
886e9c1297Now can take a VCF file as input
depristo
2010-05-18 17:06:12 +0000
e016491a3dMajor refactoring of Depth of Coverage to allow for more extensible partitions of data (now can do read group, sample, and library; in any combination; adding more is fairly easy). Changed the by-gene code to use clones of stats objects, rather than munging the interval DoCs. (Fix for Avinash. Who, hilariously, thinks my name is Carl.) Added sorting methods to ensure static ordering of header and body fields.
chartl
2010-05-18 16:58:13 +0000
2a212c497fminor improvements and bug fixes
depristo
2010-05-18 15:54:42 +0000
43544cfdf9remote control of IGV to jump to any number of loci in a file and screenshot the locus to a file
depristo
2010-05-18 12:59:36 +0000
3c022e4b0cImproved command-line-arg validation at startup.
weisburd
2010-05-18 02:46:17 +0000
04e14ef85aRefactored so it could be used for knownGene and CCDS as well as refGene
weisburd
2010-05-18 02:45:11 +0000
35b4bba35eRefactored so it could be used for knownGene and CCDS as well as refGene
weisburd
2010-05-18 02:44:10 +0000
68719615beFor multiple matches, shifted counter to be 1-based
weisburd
2010-05-18 02:41:50 +0000
73e2e32837Fix typo.
hanna
2010-05-17 21:04:00 +0000
ebd0fabf86First pass updates to annotations to work with indels. HomopolymerRun indel behavior is currently turned off by a global boolean until it's ready to go live.
chartl
2010-05-17 21:02:13 +0000
0791beab8fChecking in downsampling iterator alongside LocusIteratorByState, and removing the reference implementation. Also implemented a heap size monitor that can be used to programmatically report the current heap size.
hanna
2010-05-17 21:00:44 +0000
b7d21627abChanges to DepthOfCoverage (JIRA items) and added back an integration test to cover it. Alterations to the design file generator to output all transcripts (rather than choosing one at random).
chartl
2010-05-17 17:23:00 +0000
4235164359Removed the confusionMatrix column (of *course* this is a confusion matrix... what else would it be?!).
kiran
2010-05-14 21:55:37 +0000
6efd05831bEncapsulating annotation decoding function in order to use same fixed random seed in both VariantOptimizer and ApplyVariantClusters
rpoplin
2010-05-14 20:03:38 +0000
32389dc0a9Fixed GQ estimate when chosen genotype isn't the most likely according to the GLs.
ebanks
2010-05-14 19:17:46 +0000
1538dc0144optimizer now uses -an arguments instead of exclude and force for clarity. command-line length reduced by 50%
depristo
2010-05-14 15:41:44 +0000
88bd7a2045Reenabling UG parallelization performance tests.
hanna
2010-05-13 16:28:08 +0000
0490909285Fixed epic generic paths fail.
hanna
2010-05-13 15:59:57 +0000
7ef87e5126An integration test based on validating pileup to test parallelism in reads, reference, and RODs. This test runs in less than a minute and fell over instantly in the case of the Tribble parallelism issue.
hanna
2010-05-13 15:40:43 +0000
ceec525420Got rid of stray unicode characters in copyright message.
hanna
2010-05-13 14:47:39 +0000
aaron
depristo
kshakir
ebanks
chartl
delangel
rpoplin
weisburd
hanna
kiran