cc8d8eaedbNow that we always reserve space for two read ends when collecting stats stratified by libraries, we need to check that the second end was indeed present; otherwise the pointer is null and this was causing an exception
asivache
2010-06-28 16:40:16 +0000
26ef1f84bfUpdates to not depend on an environment variable to figure out where libraries are (helpful for installation at the Sanger).
kiran
2010-06-28 15:46:41 +0000
9a24598a98By default, don't clean reads with mates mapped to other chromosomes
ebanks
2010-06-28 15:14:20 +0000
b2127e59c4A first draft of scripts and LaTeX templates required to automatically generate slides for the 1,000 Genomes Automated Data Processing Report.
kiran
2010-06-28 14:30:50 +0000
bf5cbad04cMake the target creator a rod walker (that allows reads) so that we can easily trigger the cleaner on only known indel sites. Adding an integration test to cover this case.
ebanks
2010-06-28 13:28:37 +0000
464ac63a22Allowing N's in ALT field
ebanks
2010-06-28 11:41:32 +0000
3a9d426ca8Added hasPileupBeenDownsampled() boolean to ReadBackedPileup, so that a pileup can report whether or not (but not how much) it's been downsampled.
hanna
2010-06-28 04:56:33 +0000
8e848ccd84SAMFileWriters can now write to /dev/null without throwing exceptions, so we can remove the try/catch blocks.
ebanks
2010-06-27 03:59:10 +0000
09ccdf83b2fixing a broken test
aaron
2010-06-25 21:59:00 +0000
d6cbe4d0adBug fixes to support haploid genotypes, optimization for indexing, now tracks the line of the VCF and catches errors to tell you the line no and line when a parsing error occurred.
depristo
2010-06-25 21:08:41 +0000
5f8a3f95efThe GT field once again reigns supreme (it must be the first genotype field). Thanks for the catch Eric.
aaron
2010-06-25 21:03:05 +0000
894ad354faFixed typo in the name of the shell directory.
kshakir
2010-06-25 20:59:40 +0000
75c98c42b8Started path of deprecation of Sting's @Argument by splitting the annotation into @Output and @Input. Anything that's not an @Output should be an @Input. Checked in example qscripts that are basically todo integration tests. Replaced use of queue @Input/@Output with Sting's new @Input/@Output. This means you'll now have to doc-ument the annotations. More work on dependency resolution cycles being created in the graph during scatter/gather. Filtering nulls to avoid NPE exceptions in scala's 'Collection'.hashCode.
kshakir
2010-06-25 20:51:13 +0000
147ba68441Fixed bug with mrnaCoord field - made it count exon positions only, rather than introns & exons
weisburd
2010-06-25 19:53:32 +0000
ce27ed0d60Added missing @ClassType to memory limits.
kshakir
2010-06-25 18:56:35 +0000
dff4c06763Rev'ing Tribble with a special version that has excluded VCF 3.3
aaron
2010-06-25 18:20:51 +0000
d3848745abmoving VCF 3.3 back into the GATK so Guillermo can make changes for VCF 4 output
aaron
2010-06-25 18:20:06 +0000
b3edb7dc08two fixes for the VCF 4 parser:
aaron
2010-06-25 18:01:23 +0000
e15fe6858eDisabling test - Will need to update big-tables soon.. will re-enable after updating md5
weisburd
2010-06-25 15:43:41 +0000
efa60e5de5and add changes to the vcf used in testing
aaron
2010-06-25 02:56:02 +0000
f9c7803d4ethis got left off my last commit
aaron
2010-06-25 02:42:44 +0000
3cd0570c1eNow can run with multiple processes, multiple threads, or both
weisburd
2010-06-24 23:25:01 +0000
dae3ce2c0fchanged log dir
weisburd
2010-06-24 23:08:13 +0000
fea8054e9eUpdated long name for -l to --run-locally
weisburd
2010-06-24 22:26:45 +0000
682f9b46c6Two fixes together: 1) Some improvements to the VCF4 parsing, including disabling validation. 2) Reimplemented RefSeq in the new Tribble-style rod system.
aaron
2010-06-24 22:17:03 +0000
72e669538eUpdated arg description for -s
weisburd
2010-06-24 22:04:01 +0000
62bc7651a8fix for PSPW with DbSNP mask. Added an integration test for this case.
aaron
2010-06-24 19:31:32 +0000
4840ef6d3eAnother rev of picard for /dev/null writing.
hanna
2010-06-24 19:22:47 +0000
c32f9d78aeRev picard again, this time for error writing to /dev/null.
hanna
2010-06-24 04:08:26 +0000
bcab0eba01This replaces tearsheet.r, neatens up graphics, and allows the script to be used in R's interactive environment
corin
2010-06-24 01:02:58 +0000
8a9b2f4256removing the GLF ROD.
aaron
2010-06-23 22:51:45 +0000
17d2043354bug fix: now contigs not present in the sequence dictionary are registered properly and do not cause the script to break
asivache
2010-06-23 20:59:38 +0000
611d834092a couple of VCF 4 improvements:
aaron
2010-06-23 20:21:43 +0000
54ae0b8e4esome updates to tribble for the svn commit that will follow
aaron
2010-06-23 20:20:07 +0000
f0fc34bb8eBug fix: N's are allowed in the ref so don't fail when e.g. dbsnp has an N!
ebanks
2010-06-23 17:49:14 +0000
b6bceb39b0Fixing up output for performance tests
ebanks
2010-06-23 17:00:17 +0000
75d4736600Committing changes to comp overlap for indels. Passes all integration tests; minor changes to MVC walker.
chartl
2010-06-23 15:49:13 +0000
9b8775180eTurn on the memory improvement by default (assume the target interval list is sorted, since it is 99.9% of the time). Make the user throw a flag when it's specfically not sorted.
ebanks
2010-06-23 15:44:55 +0000
26d51bbe14Another round of optimizations from Alec. Switching the header merger to an IdentityHashMap provides another 10x+ performance boost over his previous optimization for us.
hanna
2010-06-23 14:54:58 +0000
003dd4de3eRev Picard with performance enhancements.
hanna
2010-06-22 22:54:23 +0000
0cafd3d642clip VCF alleles for indels: only a single left base, and as many right bases as align before converting to variant context.
aaron
2010-06-22 22:42:38 +0000
9872b65803clip to the null allele on the reference string in VCF 4, instead of stopping to perserve one reference base.
aaron
2010-06-22 20:52:19 +0000
b5df2705c9-Remove Nway output option -Remove in-memory sorting -Default to name-sorting (although we allow coordinate sorting with the --sortInCoordinateOrderEvenThoughItIsHighlyUnsafe flag).
ebanks
2010-06-22 20:31:55 +0000
30cf78fdc0Refactoring for a first version of scatter gather api with basic shell script implementations. Modified build script so that queue is cleaned during "ant clean".
kshakir
2010-06-22 18:39:20 +0000
18f62a346dfixing the tests; Bamboo captured the failure in the logs correctly.
aaron
2010-06-22 18:38:03 +0000
a6d3e4bd47Add code to allow reference alleles with 'N' in VariantContext, but not in the alternate allele(s). Also more updates to the VCF 4 code (fixed parsing for files without genotypes).
aaron
2010-06-22 18:26:37 +0000
4727bcda24Removing Beagle output from UG. Use ProduceBeagleInput walker instead (since it can be run post-filtration and respects the FILTER column).
ebanks
2010-06-22 16:56:37 +0000
55abfcc52fadding file my VCF4 tests depend on
aaron
2010-06-22 13:39:20 +0000
5b87a00a5fupdating with associated Tribble changes
aaron
2010-06-22 07:54:15 +0000
32f324a009incremental changes to the VCF4 codec, including allele clipping down to the minimum reference allele; adding unit testing for certain aspects of the parsing. Not ready for prime-time yet.
aaron
2010-06-22 06:31:05 +0000
de9f1f575fFixing command line parsing to accept negative number arguments. Command line definitions must now start with a letter or underscore; previously, they could start with a digit.
bthomas
2010-06-21 21:54:31 +0000
9d6a341d15Fixing the error messages thrown with bad interval arguments. I simplified the exception handling and made the messages more verbose.
bthomas
2010-06-21 21:49:41 +0000
300a18b85fUpdating the way reference data is processed, so GATK creates the .fasta.fai and .dict files automatically. If either (or both) don't exist, GATK will create them in the same folder as the fasta file. If it can't write the file, GATK will fail with a message to create them manually.
bthomas
2010-06-21 21:42:42 +0000
df1cadc4c9Fix NullPointerException when priority list is left out
ebanks
2010-06-21 13:46:54 +0000
c806ffba5fSwitching over DownsamplingLocusIteratorByState -> LocusIteratorByState. Some operations will not be as fast as they could be because the workflow is currently merge sam records (sharding) -> split sam records (LocusIteratorByState) -> merge records (LocusIteraotorByState) -> split records (StratifiedAlignmentContext), but this will be fixed when StratifiedAlignmentContext is updated to take advantage of the new functionality in ReadBackedPileup.
hanna
2010-06-21 02:11:42 +0000
1d50fc7087Misc bug fixes: fix tracking of nInsertions with sample-split pileup constructor. Fix performance issue building up pileups from pileups of individual sample data.
hanna
2010-06-20 20:32:27 +0000
f18ac069e2A refactoring / unification of ReadBackedPileup and ReadBackedExtendedEventPileup. Provides a cleaner interface with extended events inheriting all of the basic RBP functionality. Implementation is still slightly messy, but should allow users to provide separate implementations of methods for sample split pileups and unsplit pileups for efficiency's sake. Methods not covered by unit/integration tests have not been sufficiently tested yet. Unit tests will follow this week.
hanna
2010-06-20 04:42:26 +0000
57a13805daGATK now uses a optimized indexing scheme in Tribble. 5x or more performance gain on files with many genotypes. Updated integrationtest that was failing and was clearly wrong. DB=; isn't a valid annotation.
depristo
2010-06-19 21:36:41 +0000
8ff93f77e6Added evaluation module to count functional classes (missense, nonsense, etc.). At the moment, it only understands Cancer's MAF annotations. Added integration test for the functional class counting. Added better description for VariantEval.
kiran
2010-06-18 21:51:40 +0000
ae88630d52This script produces tearsheet and data processing report figures and tables when given Squid and Firehose produced data
corin
2010-06-18 21:36:29 +0000
1e06d2bf68Initial HLA Caller integration tests. Kind of painful, but will improve with code refactoring. This baby is now officially ours.
ebanks
2010-06-18 20:35:27 +0000
f44d8b150fMendelian Violation Classifier now filters violations on the fly via command line arguments; and closes unterminated homozygous regions at the end of a chromosome (so we see arms falling off in the file, rather than in the log)
chartl
2010-06-18 19:32:24 +0000
aa1852575eAdd -noVerbose flag to stop output of INFO data. Cuts runtime by 30% and output from 65Mb to 1Kb.
ebanks
2010-06-18 18:53:35 +0000
724affc3ccMajor bug fixes for the Variant Recalibrator. Covariance matrix values are now allowed to be negative. When probabilities are multiplied together the calculation is done in log space, normalized, then converted back to real valued probabilities. Clustering weights have been changed to only use HapMap and by-1000genomes sites. The -nI argument was removed and now clustering simply runs until convergence. Test cases seem to work best when using just two annotations (QD and SB). More changes are in the works and are being evaluated. Misc fixes to walkers that use RScript due to CentOS changes.
rpoplin
2010-06-18 17:37:11 +0000
c3434493b0fixed integration test for VCF Header changes.
aaron
2010-06-18 16:31:48 +0000
52477bd9e6Add some missing methods to the pileup architecture.
hanna
2010-06-18 15:03:08 +0000
5050b19457We're unable to make the naive deduper more worldly, so we're killing it instead.
hanna
2010-06-18 13:54:27 +0000
42e7ff4f28forgot to update a test, the md5sum of the underlying file changed (which is recorded in the ROD tests).
aaron
2010-06-18 13:27:56 +0000
a2c266bda3This script accpets file paths to analysis metrics tables and produces tearsheet data and data processing report graphs
corin
2010-06-18 13:02:25 +0000
b978d5946badding changes for VCF 4, mostly in the way we handle VCF headers. The header fields are now aware of the differences between different VCF formats. There was also a bunch of clean-up of out-of-spec VCF used in the tests (mismatched VCF file format fields, etc), and updates to the associated integration tests. Also some logging statements for BTI.
aaron
2010-06-18 08:23:23 +0000
32f6781ac7updating tribble with the VCF header changes
aaron
2010-06-18 08:20:44 +0000
e26a273ef5Turned the test back on
weisburd
2010-06-17 22:57:42 +0000
48cbc5ce37Merging the sharding-specific inherited classes down into the base.
hanna
2010-06-17 22:36:13 +0000
612c3fdd9dFirst pass at eliminating the old sharding system. Classes required for the original sharding system are gone where I could identify them, but hierarchies that split to support two sharding systems have not yet been taken apart. @Eric: ~4k lines.
hanna
2010-06-17 20:17:31 +0000
b694ca9633Cleanup: Don't require likelihood ROD in Beagle parameters when generating output VCF. Likelihoods file is only an input to Beagle but the Walker that generates a VCF doesn't need it, so it's silly to ask for it and it's error-prone.
delangel
2010-06-17 17:45:48 +0000
c1595a383aMore bugfixes for cases where no sample name is present.
hanna
2010-06-17 16:46:02 +0000
5972ad1199Fixes to mrl integration.
hanna
2010-06-16 20:40:10 +0000
b75ded61b8Removing obsolete rod; no longer needed given previous addition to SampleUtils. JIRA GSA-318
ebanks
2010-06-16 20:03:14 +0000
c671864228Re-allowing blacklist by read group id.
kshakir
2010-06-16 19:45:44 +0000
f003703912Allow specification of particular rods for pulling out sample names.
ebanks
2010-06-16 19:37:09 +0000
01ffa307c2When going NWay out in the cleaner, use the new *merged* header (instead of the original one) for each bam file so that it matches the new uniquified read group ids in the reads.
ebanks
2010-06-16 19:36:36 +0000
05c2f96bb4Small update to the command line docs for read_group_black_list.
kshakir
2010-06-16 19:23:34 +0000
d7f3102c3fFixed read group blacklist filter to look only at readgroups (and not the read's themselves). Otherwise, it fails when attribute tags with different meanings show up in both places (e.g. SM). Added performance improvement.
ebanks
2010-06-16 19:14:37 +0000
e77f76f8e1Reenabled downsampling by sample after basic sanity testing and fixes of the new implementation. Hard testing and performance enhancements are still pending.
hanna
2010-06-16 17:23:27 +0000
c44fd05aa1Fix for a reflection issue with generic types.
kshakir
2010-06-16 15:58:38 +0000
7a91dbd490Renamed some of the column names in Ti/Tv and Concordance modules so that they are clearer. Removed ValidationRate module (it was busted).
ebanks
2010-06-16 15:53:06 +0000
8cb16a1d45a) Cleanup, remove -input argument from BeagleOutputToVCFWalker since it's not needed. b) Added back old Beagle ROD to maintain backward compatibility (does anyone even use this???)
delangel
2010-06-16 02:13:08 +0000
d319a28be7Complete rewrite of the Beagle functionality to read from Beagle output files and produce VCF with modified genotypes. Now, a new ROD system using Tribble is in place. Beagle inputs are set using -B beagleType,Beagle,pathToBeagleFile, where beagleType can be either beagleR2, beagleLike, beaglePhased or beagleR2 (BeagleOutputToVCFWalker requires all of the above). Only pending items: -input argument is now unused and can be removed, will be cleaned later. Wiki will be updated with new usage shortly.
delangel
2010-06-16 02:01:35 +0000
asivache
kiran
ebanks
weisburd
hanna
aaron
depristo
kshakir
corin
chartl
bthomas
rpoplin
delangel
sjia