zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
5,811 Commits
1 Branch
0 Tags
257 MiB
Commit Graph

552 Commits (5dca1e4d2e44d734b95a89e46cb9e416dfd4bae0)

Author SHA1 Message Date
chartl 480859db50 Contractified version of MannWhitneyU. Some behavior has been changed: 
- Running a test when there are no observations of at least one of the sets now breaks the MWU contract
   + MWU returns Pair(Double.NaN,Double.NaN) in these instances to maintain the contract of never returning null
   + No more Double.Infinity values will appear
 - RankSumTests now probe the return values for NaNs, and don't annotate if they appear
 - For small sets where the probability is calculated recursively, the z-value is now the inversion of the error function
    and not the approximate z-value
 - UG and Annotator integration tests updated to reflect changes



git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5845 348d0f76-0448-11de-a6fe-93d51630548a
 2011-05-23 13:57:15 +00:00
depristo 6a49e8df34 Significant change to the way subsetting by sample works with monomorphic sites. Now keeps the alt allele, even if a record is AC=0 after the subset. Previously, the system dropped the alt allele, which I don't think is the right behavior. If you really want a VCF without monomorphic sites, use the option to drop monomorphic sites after subsetting. See detailed information below. 
Right now, if you select a multi-sample VCF file down (or one with filters I see) down to a smaller set of samples, and the site isn't polymorphic in that subgroup, then the alt allele is lost.  For example, when selecting down NA12878 from the OMNI, I previously received the following VCF:

1       82154   rs4477212       A       .       .       PASS    AC=0;AF=0.00;AN=2;CR=100.0;DP=0;GentrainScore=0.7826;HW=1.0     GT:GC   0/0:0.7205
1       534247  SNP1-524110     C       .       .       PASS    AC=0;AF=0.00;AN=2;CR=99.93414;DP=0;GentrainScore=0.7423;HW=1.0  GT:GC   0/0:0.6491
1       565286  SNP1-555149     C       T       .       PASS    AC=2;AF=1.00;AN=2;CR=98.8266;DP=0;GentrainScore=0.7029;HW=1.0   GT:GC   1/1:0.3471
1       569624  SNP1-559487     T       C       .       PASS    AC=2;AF=1.00;AN=2;CR=97.8022;DP=0;GentrainScore=0.8070;HW=1.0   GT:GC   1/1:0.3942

Where the first two records lost the ALT allele, because NA12878 is hom-ref at this site.  My change results in a VCF that looks like:

1       82154   rs4477212       A       G       .       PASS    AC=0;AF=0.00;AN=2;CR=100.0;DP=0;GentrainScore=0.7826;HW=1.0     GT:GC   0/0:0.7205
1       534247  SNP1-524110     C       T       .       PASS    AC=0;AF=0.00;AN=2;CR=99.93414;DP=0;GentrainScore=0.7423;HW=1.0  GT:GC   0/0:0.6491
1       565286  SNP1-555149     C       T       .       PASS    AC=2;AF=1.00;AN=2;CR=98.8266;DP=0;GentrainScore=0.7029;HW=1.0   GT:GC   1/1:0.3471
1       569624  SNP1-559487     T       C       .       PASS    AC=2;AF=1.00;AN=2;CR=97.8022;DP=0;GentrainScore=0.8070;HW=1.0   GT:GC   1/1:0.3942

The genotype remains unchanged, but the ALT allele is now preserved.  I think this is the correct behavior, as reducing samples down shouldn't change the character of the site, only the AC in the subpopulation.  This is related to the tricky issue of isPolymorphic() vs. isVariant().  

isVariant => is there an ALT allele?
isPolymorphic => is some sample non-ref in the samples?

In part this is complicated as the semantics of sites-only VCFs, where ALT = . is used to mean not-polymorphic.  Unfortunately, I just don't think there's a consistent convention right now, but it might be worth at some point to adopt a single approach to handling this.  Wiki docs updated.

Does anyone have critical infrastructure that depends on the previous convention?  Let me know so we can coordinate the change.

There's a new function subContextFromGenotypes() that also takes a Set<Allele> to handle this type of behavior.

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5832 348d0f76-0448-11de-a6fe-93d51630548a
 2011-05-21 13:59:16 +00:00
depristo e16bc2cbd9 Contracts for Java now write for GenomeLoc and GenomeLocParser. The semantics of GenomeLoc are now much clearer. It is no longer allowed to create invalid GenomeLocs -- you can only create them with well formed start, end, and contigs, with respect to the mater dictionary. Where one previously created an invalid GenomeLoc, and asked is this valid, you must now provide the raw arguments to helper functions to assess this. Providing bad arguments to GenomeLoc generates UserExceptions now. Added utilty functions contigIsInDictionary and indexIsInDictionary to help with this. 
Refactored several Interval utilties from GenomeLocParser to IntervalUtils, as one might expect they go

Removed GenomeLoc.clone() method, as this was not correctly implemented, and actually unnecessary, as GenomeLocs are immutable.  Several iterator classes have changed to remove their use of clone()

Removed misc. unnecessary imports

Disabled, temporarily, the validating pileup integration test, as it uses reads mapped to an different reference sequence for ecoli, and this now does not satisfy the contracts for GenomeLoc


git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5827 348d0f76-0448-11de-a6fe-93d51630548a
 2011-05-20 15:43:27 +00:00
rpoplin 40797f9d45 Ensuring a minimum number of variants when clustering with bad variants. Better error message when Matrix library fails to calculate inverse. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5793 348d0f76-0448-11de-a6fe-93d51630548a
 2011-05-12 01:48:37 +00:00
ebanks dfdef2d29b PLEASE READ ME! In order to prepare for the upcoming changes to VCF4, we felt it was best to split up the vcf3 and vcf4 codecs (vcf4 is not backwards compatible to vcf3 and certain changes are too complex to handle in both codecs). Using the 'VCF' rod type in the GATK will now throw a UserException for vcf3.2 or vcf3.3 files telling you to use the 'VCF3' type instead (and vice versa). Integration/unit tests have been updated. For programmers: note that there is currently a lot of code duplication in the two codecs (although I pulled out the easy stuff to a VCFCodecUtils class); however WE ARE FREEZING THE VCF3 CODEC AND WILL NO LONGER MAKE CHANGES TO IT. All updates/improvements will be targetted to the vcf4 codec only as vcf3 is there only to be able to read legacy files. People should really be using vcf4 files only. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5787 348d0f76-0448-11de-a6fe-93d51630548a
 2011-05-11 12:07:44 +00:00
ebanks 15c7bd82a5 Fix for IndelRealigner memory problem. Now the Constrained mate fixing writer is told whether a read has been modified and, if it wasn't, can dump it when the cache needs to get flushed at places with tons of coverage. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5777 348d0f76-0448-11de-a6fe-93d51630548a
 2011-05-06 19:34:41 +00:00
delangel 7d7ce6cf00 Two embarassing bug fixes: 
a) Forgot to convert from phred to log-prob when computing gap penalties from recal table.
b) Forgot to uncomment code to correctly deal with hard-clipped bases in a read. But because of this, had to do a short term workaround to at least temporarily return class from hardClipAdaptorSequence to GATKSAMRecord. Otherwise, I get exceptions when casting because somehow some reads in HiSeq get to be SAMRecord (which GATKSAMRecord inherits from) but some reads get to be BAMRecords (which can't be cast into GATKSAMRecord), not sure why.



git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5771 348d0f76-0448-11de-a6fe-93d51630548a
 2011-05-05 17:08:34 +00:00
carneiro 3882d1b9c0 fixing the build \o/ 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5767 348d0f76-0448-11de-a6fe-93d51630548a
 2011-05-05 00:57:49 +00:00
rpoplin 6c7a0adc76 Updating VariantGaussianMixtureModelUnitTest to use truth sensitivity cutting 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5750 348d0f76-0448-11de-a6fe-93d51630548a
 2011-05-04 13:56:01 +00:00
rpoplin 23cd3a7a5d Moving VQSR v2 to core. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5740 348d0f76-0448-11de-a6fe-93d51630548a
 2011-05-03 20:20:06 +00:00
ebanks d4cbd8691c Make the default that we only output SNPs (so that when I make another release we don't get flooded with questions about why the UG is all of a sudden so slow) 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5729 348d0f76-0448-11de-a6fe-93d51630548a
 2011-05-03 16:38:55 +00:00
ebanks deed7c47a1 Continuing the epic fail, some of our existing integration tests were wrong because of the lazy loading failure. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5712 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-29 17:54:41 +00:00
ebanks ab9ffb1a74 Epic failure on the lazy loading of genotypes: if the input VCF had its samples unsorted and we used a walker that didn't require genotypes, then we would sort the samples but not load genotypes (and therefore the genotypes wouldn't match the samples anymore). Added simple integration test to cover this case. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5711 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-29 16:03:45 +00:00
rpoplin b7334dcc1e Rank sum test annotations are the Z-scores from the test instead of the p-value. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5707 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-28 14:35:00 +00:00
ebanks 45081c32d7 continuing from last night, the integration tests weren't covering the right behavior either 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5706 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-28 13:30:57 +00:00
delangel 600617a63c Enabled code to deal with hard-clipping adaptor sequence when processing reads in pileup in indel caller. Proven now that changes are minimal (4 less calls in NA12878 chr20, quals slightly different), minor changes in vcf fields in integration tests. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5679 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-22 14:10:33 +00:00
ebanks 49ea07acce My fixes to Tribble yesterday revealed that some of the test VCFs for integration tests were actually malformed. Also, Guillermo updated the b37 dbSNP VCF and that broke some tests. Should be good for now. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5655 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-17 03:39:11 +00:00
depristo 8ed9c0f518 VariantsToTable now blows up by default if you ask for a field that isn't present in a record. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5636 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-14 14:42:43 +00:00
rpoplin 30a19a00fe Fix for when running with EMIT_ALL_SITES but not GENOTYPE_GIVEN_ALLELES. Still want to emit a site even when over the deletion fraction for example. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5617 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-12 20:07:06 +00:00
delangel 3b424fd74d Enable new indel likelihood model by default, cleanup code, remove dead arguments, still more cleanups to follow. This isn't final version but at least it performs better in all cases than previous Dindel-based version, so no reason to keep old one around. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5615 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-12 17:54:46 +00:00
ebanks b6e7b5dace Updating to reflect my recent Tribble fix 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5601 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-08 11:48:00 +00:00
ebanks cd61ef7169 Re-enabling multi-threaded integration tests. To make this work, downsampling and annotations are disabled for this test so that we don't have randomization issues for it based on which shards get executed first. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5597 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-08 03:07:39 +00:00
ebanks af09170167 As I threatened yesterday, I've moved the various and disparate randomization code out of the walkers. Now they all (except VQSRv1, whose days are numbered anyways) use a static generator available in the engine itself. Please use this from now on. The seed is reset before every individual integration test is run. I think there may still be an issue with the IndelRealigner but I need to confirm with the commit to see what testNG does. Integration tests are already broken anyways, so no big deal. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5589 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-07 17:03:48 +00:00
rpoplin 3f3f35dea0 UnifiedGenotyper now BAQs via ADD_TAG to facilitate using BAQed quals for GL calculations but unBAQed quals for annotation calculations. UnifiedGenotyper now produces SNP and indel calls simultaneously. 40 base mismatch intrinsic filter removed from UG to greatly simplify the code. RankSumTests are now standard annotations but the integration tests are commented out pending changes that will allow random annotations to work. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5585 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-06 19:06:24 +00:00
ebanks 4b451314b2 Only store a read in the mate hash if it could possibly be moved. This reduces memory consumption especially when dealing with a case of tons of unmapped reads at the end of the bam; however, it's only mildly helpful for chr1 of the Papuans (there's a truly massive pileup 120Mb into it; more thought needed at a later point). Integration tests changed only because some of the reads in the original bam were busted to begin with (it's an old pilot 1000G bam). 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5580 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-05 22:20:09 +00:00
depristo 095125152b Updated to now longer include 2nd-best base output 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5567 348d0f76-0448-11de-a6fe-93d51630548a
 2011-04-03 20:13:10 +00:00
depristo 3bcd4c5d75 --simplifyBAM is now in the SAMFileWriterArgumentTypeDescriptor, as suggested by map. PrintReads has an integrationtest now that writes out a 1 MB bit of HiSeq normally, with compress 0, and with simplifyBAM on. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5521 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-26 14:57:18 +00:00
ebanks 69646ff840 ... and the corresponding integration test update 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5496 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-23 01:58:07 +00:00
ebanks 1c95208e26 Finally found the bug that everyone is reporting on GS. Iterators on PriorityQueues aren't guaranteed to return elements in sorted order (a pretty stupid contract) - so we were passing items to the constrained writer out of order. Just do a Collections.sort instead (1 line of code). Happy father's day! 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5476 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-18 21:28:19 +00:00
depristo 3e3ec85807 Checked for consistency with the previous integration tests, and updated the walker and test to use the new I/O system (always prints 4 digits on floats. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5433 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-13 15:24:22 +00:00
depristo ee8f2871f7 A better output for Genotype Concordance summary. Now does only % comp hom-ref called hom-ref, het called het, and hom-var called hom-var, which are the quantities we typically show in slides. Updated intergration tests to reflect this change. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5429 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-12 02:03:48 +00:00
ebanks 3596c56602 New attempt at the constrained movement version of the indel realigner (I've kept around the old writer for now). The new contract is that the realigner must ask permission before trying to clean an area; permission will be denied by the CM-Manager if it was required to flush its cache of reads because of too much depth within a distance of maxInsertSizeForMovingReadPairs. Added integration tests to cover different max cache sizes, including an expected exception when too small a value is chosen. The actual logic changes were fairly minor - much of this commit is really just some cleanup. I'd like to throw 1000G Phase I at it, but will respectfully wait for Ryan to hit his deadline before doing so. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5414 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-10 02:48:29 +00:00
rpoplin ff7edc4493 Minor bug fix in empiricalMu prior calculation in VQSR. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5412 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-10 00:42:38 +00:00
rpoplin 509daac9f7 Minor bug fix in k-means implementation. Updating VQSR integration tests in preparation for VQSRv2 by removing some unused features such as VariantDatum.weight and ti/tv cutting. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5410 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-09 00:26:28 +00:00
delangel 00ac51acc8 Added several integration tests for UG indel caller: 
- Basic
- Multiple technology
- Test minIndelCnt parameter

Added also 2 disabled tests:
- Parallelization: issue w/code right now is that if -nt > 1, filter field shows "PASS" instead to ".", cause TBD
- Genotype given alleles mode: code not working yet.



git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5404 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-08 16:21:21 +00:00
kiran d0598c7a04 Somehow missed this test when I was updating the md5s 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5400 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-07 23:53:42 +00:00
kiran b6339967f8 Updated GenomicAnnotator integration tests to include the -NO_HEADER argument so that they tests op yelling about trtrivial differences 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5398 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-07 23:07:01 +00:00
kiran 43056d0188 Fixed integration test to reflect changes regarding when comp tracks got subset to fewer samples and whether no-call sites would get pulled in for comp tracks. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5393 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-07 20:25:57 +00:00
kshakir dc33fbed7c Switched the CVUnitTest broken info from an Integer to a String since as of r5383 Integers are no longer broken when converted to Floats. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5390 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-07 16:33:14 +00:00
depristo af71576a07 CalculateChromosomeCounts() now only calculates AC, AF, and AN when there are genotypes. Can now combine variants with headers that differ in only whether a field is a integer or a float. Updated CombineVariants integrationtest, as incorrect AC values where being calculated in the previous GS outputs. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5383 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-06 19:25:52 +00:00
kiran 1861ca90fc A change to the definition of CpG sites (is now, from 5' to 3' a CG dinucleotide in the reference, and the CpG site is at the C, rather than either at the C or a G). 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5373 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-04 15:36:07 +00:00
ebanks bb969cd3a2 EMIT_ALL_SITES now does exactly that - even when there's no coverage or too many deletions 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5343 348d0f76-0448-11de-a6fe-93d51630548a
 2011-03-01 05:05:00 +00:00
ebanks 5ac9af472c Adding performance test for case with very high coverage (> 600,000x) over an interval 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5336 348d0f76-0448-11de-a6fe-93d51630548a
 2011-02-28 19:48:56 +00:00
ebanks cba88a8861 Elegant solution to the determinism problem: force testNG to run tests in the order that I want it to. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5312 348d0f76-0448-11de-a6fe-93d51630548a
 2011-02-24 21:32:35 +00:00
ebanks 15dfac6bf7 Updating integration test to be in sync with previous commit 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5309 348d0f76-0448-11de-a6fe-93d51630548a
 2011-02-24 20:21:58 +00:00
ebanks 06e3c34e7f Updating performance test to be in sync with previous commit 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5308 348d0f76-0448-11de-a6fe-93d51630548a
 2011-02-24 20:13:35 +00:00
chartl 97e1a5262e -ct x no longer includes coverage in the previous bin 
BatchMerge - additional support for indels (can't just test the alternate allele when it's an extended event, must also specify that you want to use the dindel model when you actually test the allele)



git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5300 348d0f76-0448-11de-a6fe-93d51630548a
 2011-02-24 15:52:04 +00:00
ebanks ee6f112556 Phase 3: constrained movement is now the only option available in the realigner (so I guess technically it's not really an option). Several command-line options are deprecated. Code cleaned up. Wiki updated. Release coming. One phase left... 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5299 348d0f76-0448-11de-a6fe-93d51630548a
 2011-02-24 14:59:48 +00:00
ebanks 93888e570b Phase 2: after hours of testing, confirming that constrained mode looks good so moving the integration tests over to use it. Some cleanup. More cleanup coming in Phase 3. 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5298 348d0f76-0448-11de-a6fe-93d51630548a
 2011-02-24 06:23:41 +00:00
ebanks c59c8b9872 Phase I of my promise to Mark: fleshed out integration tests for Indel Realigner 
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5297 348d0f76-0448-11de-a6fe-93d51630548a
 2011-02-23 21:05:20 +00:00