88ca1fb22cLazy loading reflections so Queue can hack the classpath before the PluginManager looks for classes. Removed extra quotes from 'cd' pre-exec command.
kshakir
2010-08-19 20:29:52 +0000
0028b884d8Reformatting and tweaks to the end-to-end pipeline
chartl
2010-08-19 20:29:48 +0000
63ada20da5allow RefSeq files to optionally contain the header line, which is the default output from the UCSC table browser
aaron
2010-08-19 20:25:37 +0000
effeedf1a3Updated Bayesian phasing method to output per-site phasing statistics (and to not cap PQ at 40)
fromer
2010-08-19 19:55:47 +0000
04e5b28f6dupdates for VCF; we can no longer cache genotypes or alleles in a static array, this is bad for sharred memory parallel runs. One instance per codec was better for performance than using ThreadLocal code.
aaron
2010-08-19 19:34:44 +0000
8054b6b295Changing a name of a column for variantevals output for easier reading by R--let me know if this needs to be updated elsewhere; it's just a space to an underscore.
corin
2010-08-19 19:18:16 +0000
4b94f8c21bSilly me, I forgot to check for the contig boundaries. Thank goodness for performance tests!
ebanks
2010-08-19 18:40:26 +0000
f16bb1e830fix for a bug in package utils.
aaron
2010-08-19 15:01:50 +0000
15c5aa6e48Efficient iteration over all possible combinations of variable assignments, for variables of arbitrary cardinalities
fromer
2010-08-19 14:14:37 +0000
1ec305cd15Fix for running the cleaner at the lane-level for known indels only: instead of relying on the reads to get the reference sequence, we now use an IndexedFastaSequenceFile in all cases and pad the reference with bases on either end. This allows us to deal with cases in which we are trying to clean just a single deletion-containing read with tiny LOD (so the read needs to be pushed off the seen reference; @Reference doesn't yet work for Read Walkers) and has the added benefit of allowing us now to get much larger known indels that aren't completely covered with reads. Thanks to Matt for the advice.
ebanks
2010-08-19 11:31:13 +0000
98f7679619Fixed the bug reported on GS regarding a clipped read that got moved several hundred bases away. The code that got triggered here was written back in the original version of the cleaner and it never actually did the right thing.
ebanks
2010-08-19 01:42:48 +0000
3dc4d3c3a9removing the custom reflections library from the libs, and adding a release version. Hopefully this will fix the problem Menachem has been seeing with random JVM crashes. Also removed the auto-deletion of the reflections jar, and removed the very old OmniPlan document we had checked-in.
aaron
2010-08-19 00:42:37 +0000
1c3f2be829Change for Chris: don't depend on other Perl modules.
kiran
2010-08-17 22:19:13 +0000
cb74806bb5Enabling support for any scala used in QScripts.
kshakir
2010-08-17 22:05:56 +0000
8f15b2ba72Memory optimization for the VariantRecalibrator. Only add variants to the list if they pass the novelty and qual filters.
rpoplin
2010-08-17 21:57:28 +0000
b7c60b9729Queue now uses its own version instead of the gatk version. Added a Queue release directory.
kshakir
2010-08-17 19:34:23 +0000
ede87a03c2Nicer plotting routine for tranches. Add a third arg to suppress the legend.
depristo
2010-08-17 19:20:58 +0000
618c69f8dcMore updates to the CleanBamFile pipeline. Added the a CommandLineFunction.jobDependencies that will explicitly force a function to wait for a file, even if the value isn't otherwise listed on an @Input. More bug fixes and refactoring of functions.
kshakir
2010-08-17 14:59:42 +0000
e632d9b83dremove some dependencies on out of date methods from the tests
aaron
2010-08-17 00:07:26 +0000
c1df293febremove testing code from tribble track builder, set the command line program in walker test to null to reclaim memory in integration tests, and removed some orphaned intergration tests.
aaron
2010-08-16 23:52:01 +0000
3a4977c75eRe-add the 1KG trigger as a comp as well
chartl
2010-08-16 18:19:47 +0000
578e7fa36dDon't output -0 as qual value in VariantRecalibrator.
rpoplin
2010-08-16 16:47:58 +0000
3d63302b70Deprecated. Use SelectVariants instead.
kiran
2010-08-16 15:07:50 +0000
20db00a3e8Lazy reference loading; the engine doesn't fetch the reference bases until you actually call ref.getBases(). With the new hidden --dontUpdateUG to table recalibrator this is 2-3x faster than before. Enabled for locus, read, and rod walkers.
depristo
2010-08-16 13:46:22 +0000
9ab647b730adding checks to the RefSeq rod for line's that contain less than the required number of columns (we expect there to be 16 columns)
aaron
2010-08-16 13:34:32 +0000
cc58a27b00fix for broken unit test; make sure when we can't get an index off of disk, the internal method returns null
aaron
2010-08-16 13:12:32 +0000
b23545fafare-enable the check for up-to-date versions in the Tribble index.
aaron
2010-08-16 12:47:58 +0000
37586d3a43Don't exception out when bad aligners emit wonky alignments; instead, just don't clean
ebanks
2010-08-16 02:36:04 +0000
a36951f11a@output and @input arguments for table recalibration for use with Q
depristo
2010-08-14 18:36:28 +0000
61064d7075GenotypeConcordance log file -- if provided, GC module will write FN/FP information to this file by context
depristo
2010-08-14 18:35:57 +0000
0d209d5442Nicer printing out of clustering
depristo
2010-08-14 16:02:13 +0000
307c8ca027Created a new playground script for cleaning bams in Firehose. Some refactoring of Queue extensions for reusability in scripts. Putting the extensions into the Queue.jar after building them. More updates to GATK walker arguments specifying @Input and @Output for Queue.
kshakir
2010-08-13 23:52:24 +0000
dfe2922b5eFirst working version of statistical haplotype phaser
fromer
2010-08-13 21:29:45 +0000
f36c0ed613Stop building obsolete VCFTools and CGUtilities
ebanks
2010-08-13 19:28:36 +0000
8e46d5de04Printing to INFO where to find the job output files.
kshakir
2010-08-13 19:26:53 +0000
222f61df87Bug fix for damoskow in TableRecalibration. Shouldn't try to update the reference mismatch rate tag for an unmapped read.
rpoplin
2010-08-13 18:57:07 +0000
80a70ccf03Repopulating rodsToSamples. Code reviewed by Eric.
kshakir
2010-08-13 17:07:18 +0000
cb144734c0Getting rid of GenotypeWriter interface. Of note: - GATKVCFWriter deleted, to be replaced if absolutely necessary when VCF writing goes into Tribble. - VCFWriter is now an interface, for easier redirection. - VCFWriterImpl fleshes out the VCFWriter interface.
hanna
2010-08-13 16:33:22 +0000
542d394e09Cleaning up Queue debugging output. -l DEBUG with local programs now prints out the stdout/stderr of the programs as they are run. More documentation in the examples with a new even simpler CountReads example. Took out unused option to build Queue GATK extensions separately.
kshakir
2010-08-13 15:54:08 +0000
49a3db9dfeA brief implementation of a QD calculation that is not quite so bimodal for known variants (multiplicatively penalizes QD by (n variant samples)/(n variant alleles) ). Not sure how helpful this will be (which is why it is in oneoffs). Seems nice on MCKD1, but I'm still playing with the optimization.
chartl
2010-08-13 15:42:37 +0000
c6a8fba922Occasionally if a JEXL expression results in no variants being captured (like "QD > 20.0" on filtered variants) the per-sample mapping from samples to eval objects can be empty. This semi-hacky fix prevents null pointer exceptions in setting up the resulting empty table (by jumping straight to it in this case)
chartl
2010-08-13 15:37:45 +0000
f874e548aaShame on us. FlagStat used ints instead of longs, so we ended up getting negative read counts
ebanks
2010-08-13 03:00:57 +0000
71c4d3f33dMoving pointer to b36 reference from /broad/1KG to /humgen/1kg
ebanks
2010-08-13 00:54:34 +0000
162febdef8Added Queue packages, which must be run with 'ant queue package'. To assist with the above no longer removing jars during a new build, so 'ant queue dist' will still have the Queue.jar.
kshakir
2010-08-12 17:45:53 +0000
25a23218c6Trying a build server fix via google: only running the ivy taskdef once.
kshakir
2010-08-12 05:10:26 +0000
cd5d42618fKilling old version of ivy jar.
kshakir
2010-08-12 01:32:03 +0000
f39dce1082Exposed CommandLineFunction defaults to the Queue.jar command line (see -help). Added ability to skip up-to-date jobs where the outputs are older than the inputs. Changed -T CountDuplicates --quiet to --quietLocus so that Queue GATK extensions can use both short and full argument names. Short names can be used to set values on Queue GATK extensions, for example: vf.XL :+= myFile Moved Hidden from the GATK to StingUtils. Updated ivy from 2.0.0 to 2.2.0-rc1 to fix sha1 issue: http://bit.ly/aX72w7 Added Queue to javadoc and testing build targets. Added first Queue unit test. Another pass at avoiding cycles in the DAG thanks to all function I/O being files.
kshakir
2010-08-11 21:58:26 +0000
8c08f479231) Make sure that the table size is set correctly in finalize() 2) Make sure variants are biallelic before asking for isTransversion()
chartl
2010-08-11 20:32:22 +0000
41d57b7139Massive cleanup of read filtering. - Eliminate reduncancy of filter application. - Track filter metrics per-shard to facitate per merging. - Flatten counting iterator hierarchy for easier debugging. - Rename Reads class to ReadProperties and track it outside of the Sting iterators. Note: because shards are currently tied so closely to reads and not the merged triplet of <reads,ref,RODs>, the metrics classes are managed by the SAMDataSource when they should be managed by something more general. For now, we're hacking the reads data source to manage the metrics; in the future, something more general should manage the metrics classes.
hanna
2010-08-11 20:17:11 +0000
86bd55408eno INFO output now that it's the default
ebanks
2010-08-11 18:56:28 +0000
7385cce494Useful tool for calculating the perentage of misaligned reads at homozygous non-ref indel sites
ebanks
2010-08-11 17:57:44 +0000
cc9e6b4ad9Moved into Tribble to be with VC
ebanks
2010-08-11 17:14:32 +0000
14e492fa80fix for a problem in readNextRecord() of BFS, where we'd go looking for the next record far into in the next contig because (f.getEnd() >= start) was never true once we cycled to a new conitg. Added a check for contig identity. Also, removed duplicate HW calculation classes in the GATK and Tribble.
aaron
2010-08-11 17:01:38 +0000
e0abb73fd7plot now assumes 1 / 1000 is the min error rate, not 1/100
depristo
2010-08-11 14:48:22 +0000
6037443e55Handle interactive and non-interactive modes more elegantly.
kiran
2010-08-11 02:38:53 +0000
a7409df1a6Be more robust to missing or empty files in VariantEval output.
kiran
2010-08-11 02:22:50 +0000
23b5d71e76A quickly hacked together replacement for AnnotateVCFwithMAF.py, which doesn't work anymore with Cancer's updated annotator. Takes an annotated MAF file and imports the annotations into the VCF file. For the MAF annotator's DNP and TNP annotations (which I think are likely to not be correct, given the lack of phasing information or even proper association to the same sample), just propagate the annotation from the previous annotated variant to which the multinucleotide polymorphism was associated.
kiran
2010-08-11 00:20:08 +0000
cd4cd6db81Added option to print out discordant sites in GenotypeConcordance
flannick
2010-08-10 19:55:19 +0000
18fc5c8c3eInitial implementation of annotator to compute allele balance for each sample
flannick
2010-08-10 19:40:17 +0000
1dc373b9d0Initial implementation of evaluator to compute popgen theta statistics
flannick
2010-08-10 19:36:34 +0000
0a8ebcb4f9moving tests over from the GATK to Tribble, and added a speed-up to the readNextRecord() that Mark suggested. Also removed the contained flag from the queries to Tribble in the GATK.
aaron
2010-08-10 17:54:59 +0000
3ff6e3404eAlleles are now returned in a consistent order, so we can deal with tri-allelic sites
ebanks
2010-08-10 15:21:10 +0000
67063deb16Removed coloring by mixture weight. Each cluster gets a distinct color, and the legend indicates which cluster has which id and its weight
depristo
2010-08-10 14:28:24 +0000
672bee295cnow plots tranches separately from optimizer
depristo
2010-08-10 12:02:52 +0000
cd2d051209full path to Rscript
depristo
2010-08-10 12:02:38 +0000
9b432d08011kg script now works
depristo
2010-08-10 12:01:18 +0000
d514c424fdadding tests for BTI in the ROD validation tests
aaron
2010-08-10 06:05:40 +0000
ca5b274f16Unit, integration, and performance tests are all busted, so this is a good time to make a big commit...
ebanks
2010-08-10 04:18:29 +0000
0f78f70ed4fix for feature source in Tribble; we need to check that the record coming back isn't null. Also in the GATK added code to set the default logging level in integration tests to WARN, with the default level change they were spewing a bunch of text.
aaron
2010-08-10 02:57:23 +0000
419a36f74cStarting the clean up of the sting.utils.genotype code which is all either moving to Tribble, moving to sting.utils.vcf, or being removed.
ebanks
2010-08-10 02:16:05 +0000
2a4a4b0aabVariantRecalibrator now calls plot_Tranches directly so it works on the farm
depristo
2010-08-09 23:17:16 +0000
c2c0c1f57cRemoving used --enable_overlap_filters argument; Eric assures me this won't break the currently broken tests
depristo
2010-08-09 22:27:13 +0000
0f29f2ae3ffixes for the Tree index, and some small clean-up in the GATK.
aaron
2010-08-09 20:41:50 +0000
3eee3183fdChecking in the tiger team changes. LOD calculation modified. -qScale is back in case people need it.
rpoplin
2010-08-09 20:41:03 +0000
0eeb659aa3Useful utility function to print out the Allele as a String since toString prints out * for refs. It was annoying to keep seeing new String(Allele.getBases()).
ebanks
2010-08-09 20:35:56 +0000
d0ecb8875aAdded - a class to count functional annotations by sample (currently for the MAF annotation strings, soon to be migrated to genomic annotator once it is up and running)
chartl
2010-08-09 20:09:13 +0000
5b0b9e79baprotect against nulls
aaron
2010-08-09 19:21:39 +0000
8944800f60Minor refactoring for Ryan
depristo
2010-08-09 18:05:23 +0000
c66e93d86eFixed a problem when mixing queue with other targets, such as 'ant clean oneoffs queue' and the STING_BUILD_TYPE environment is set.
kshakir
2010-08-09 17:59:51 +0000
4f51a02deaChanged logging level to default at INFO instead of WARN. Changes to StingUtils command line for use in Queue, replacing Queue's use of property files. Updates to walkers used in existing QScripts to add @Input/@Output. RMD used in @Required/@Allows now has a new default equal to "any" type. New QueueGATKExtensions.jar generator for auto wrapping walkers as Queue CommandLineFunctions. Added hooks to modify the functions that perform the Scattering and Gathering (setting their jar files, other arguments, etc.) Removed dependency on BroadCore by porting LSF job submitter to scala. Ivy now pulls down module dependencies from maven.
kshakir
2010-08-09 16:42:48 +0000
30178c05c5providing a way to specify how you'd like -BTI combined with your -L options; set BTIMR to either UNION (default) or INTERSECTION.
aaron
2010-08-09 14:00:52 +0000
f3883585d0removing the build lines I inadvertently committed. As a note: these are the lines you need to add if you want to debug tests, just don't check them in.
aaron
2010-08-09 13:48:16 +0000
9aa70d9c7cReplaced by SelectVariants
kiran
2010-08-08 07:07:42 +0000
758ab428f5Better logging info for the samples being selected and the sample expressions being ignored.
kiran
2010-08-08 07:03:37 +0000
e242a8f143Put single quotes around the regex. This isn't strictly necessary through the integration test machinery, but *is* necessary at the console, and it's convenient to be able to cut and paste this.
kiran
2010-08-08 05:56:57 +0000
13f29660bbIntegration test for SelectVariants. Tests a complex case with an explicit sample selection, sample selection by regex, exclusion of non-variant and filtered loci, and JEXL selection on low allele-frequency variants
kiran
2010-08-08 05:49:47 +0000
637a1e5055Updating to use the new VA interface
ebanks
2010-08-08 05:31:01 +0000
bd6d5a8d51Adding command-line header to VA and VF
ebanks
2010-08-08 05:21:15 +0000
64446f0ddfAvoid NaNs in the final output.
kiran
2010-08-08 05:16:52 +0000
3f6e44dc71Updated recalibrator and cleaner to output full command-lines in the bam header
ebanks
2010-08-08 04:39:18 +0000
kshakir
aaron
fromer
chartl
corin
ebanks
kiran
rpoplin
depristo
hanna
flannick