zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Commit Graph

  • 33ae256186 a start to some of the infrastructure for Tribble, including dynamic detection of new RMD; not nearly wired in or complete yet. aaron 2010-02-18 18:43:52 +0000
  • bbbad79f8c Forgot to remove debugging code ebanks 2010-02-18 18:12:58 +0000
  • cc3c18d3e1 a small dbsnp file for Tribble testing aaron 2010-02-18 18:12:55 +0000
  • 7669eaaeb3 Optimizations to the cleaner algorithm; reduce total runtime by almost 20%. ebanks 2010-02-18 18:10:56 +0000
  • 8fd3351971 adding a stripped down Tribble library for the start of integration aaron 2010-02-17 21:29:25 +0000
  • 79ab7affda - Change sortOnDisk option to sortInMemory - Fix horrible cleaner bug - Trivial optimizations to cleaner code - more significant ones coming soon. ebanks 2010-02-17 20:52:57 +0000
  • 2520889cb3 Check for bad intervals and don't emit them ebanks 2010-02-16 21:42:36 +0000
  • f859e14cc7 Allow no-call alleles to propagate through to the MAF file kiran 2010-02-16 20:46:16 +0000
  • 217deb9809 Changed the INFO field delimiter from a comma to a semicolon kiran 2010-02-16 20:44:57 +0000
  • 653f70efa2 added methods to validate an interval before you try to make a GenomeLoc: boolean validGenomeLoc(). aaron 2010-02-16 20:35:35 +0000
  • f02e94ab6f Eliminate the rescale factor -- heatmap automatically normalizes the data chartl 2010-02-16 16:34:33 +0000
  • 0e4b5ad9c6 Check to ensure sample status is "Complete" before writing out the bam file chartl 2010-02-16 15:36:42 +0000
  • 37fa1bf0cc Added heatmap function chartl 2010-02-16 15:12:54 +0000
  • 01af3d0663 Update an error message :) chartl 2010-02-15 23:24:06 +0000
  • 951b7a2433 First of what will be an increasingly useful set of tools, compiled into one command-line runnable library -- the goal is to have one plotting library that's callable because of limitations on the number of files you can package with a GenePattern module. chartl 2010-02-15 16:51:47 +0000
  • 81313d9452 added class VCFMerge jmaguire 2010-02-15 14:41:50 +0000
  • 0ef50bcae7 - update to match recent changes in the VCF parser - compute Het Error Rate in VCFConcordance - changes to the frequency-specific optimizer jmaguire 2010-02-15 14:27:01 +0000
  • 8072e9aed5 should never commit without running intergration tests. depristo 2010-02-12 23:42:37 +0000
  • a1a3d5fcb0 Support for reading in table of rsIDs -> dbSNP builds to back generate a dbSNP build X from a single file. Very useful indeed. dbSNP -> VC now captures the rsID in the context depristo 2010-02-12 22:40:55 +0000
  • 28f24ca2ae made some private member/methods protected to allow for subclassing kcibul 2010-02-12 21:16:00 +0000
  • 232d884578 Got back most of the performance lost when I fixed the dropped reads problem. hanna 2010-02-12 19:59:56 +0000
  • 04a2784bf7 Initial commit of tools under development for data QC through firehose. chartl 2010-02-12 19:13:24 +0000
  • 77af5822d4 Correcting my incomplete understanding of how the BAM file index actually works. hanna 2010-02-12 16:15:19 +0000
  • 5f74fffa02 Massive improvements to VE2 infrastructure. Now supports VCF writing of interesting sites; multiple comp and eval tracks. Eric will be taking it over and expanding functionality over the next few weeks until it's ready to replace VE1 depristo 2010-02-12 15:26:52 +0000
  • 197dd540b5 added root GATKData variable depristo 2010-02-12 15:25:34 +0000
  • c6f6948f9d Haiku: Eric is a fool. Matt found his really dumb bug. Eric is humbled. ebanks 2010-02-12 04:51:56 +0000
  • d4e4120ca1 Some useful changes that I've had laying around for a while - deletes files from failed runs, automatically adds a memory limit to java commands where one isn't specified, and touches files on the local machine after command completion to get around the problem with the times not being perfectly synchronized across LSF nodes. kiran 2010-02-12 00:18:34 +0000
  • ecebf0bc62 Bug fix for null pointer exception in AnalyzeAnnotations if -name argument isn't specified rpoplin 2010-02-11 18:39:26 +0000
  • ad608d0e9d Cleaned up documentation on SecondaryBaseTransitionTableWalker and added Read Group and Allele Balance to the info. mmelgar 2010-02-11 17:20:35 +0000
  • 34e566c90d Fixed bug where new sharding system wasn't grabbing the reads that start at the end of a bin. Caused by what I currently believe to be a bug in Picard -- will verify with Alec. hanna 2010-02-11 17:00:04 +0000
  • 96fee7cf7a Disabling input of known indels for use as alternate consenses. When we get rods in a read traversal, it will be trivial to hook it into the cleaner (the code is already there). ebanks 2010-02-11 15:52:21 +0000
  • e491b42951 Dumb little script that grabs Picard metrics (alignment, hybrid selection, insert size) from picard_aggregation given the path to the bam file; zips them up, and spits them out; for use with Firehose chartl 2010-02-11 14:09:30 +0000
  • a4a2c9b172 Deal with bad input; also N-way out isn't default. ebanks 2010-02-11 03:44:56 +0000
  • dc885ba386 Fix for some correctness bugs found during early performance testing, phase 1. hanna 2010-02-10 22:32:25 +0000
  • 60f05379a7 fix typo. Check explicitly that fingerprint files exist. chartl 2010-02-10 21:29:54 +0000
  • d16a1b5645 Simple python script for generating a firehose-parseable text file from MS-Dos formatted TSV spreadsheets (of the type that we get from project management). Will be deprecated in a few weeks with the advent of direct BSP ID entries, but useful until then. chartl 2010-02-10 20:25:40 +0000
  • c66861746a improvements to ve2, including more meaningful mendelian violation counting. Support for VCF emitted interesting sites, annotated according to the evaluations themselves. Basic intergration test for VE2 started depristo 2010-02-10 16:12:29 +0000
  • 3de72daa88 Removing an accidently added import statement. rpoplin 2010-02-10 15:54:24 +0000
  • 0b1e243a7b CountCovariates now sorts the list of standard covariate classes coming from PackageUtils.getClassesImplementingInterface(). As a result some of the integration tests now make use of -standard rpoplin 2010-02-10 15:52:20 +0000
  • 6652b992f7 The new cleaner can now use known indels to create alternate consenses for cleaning. ebanks 2010-02-10 04:39:15 +0000
  • 0250338ce7 Basic use cases for merging BAM files with the new sharding system work. hanna 2010-02-09 22:14:37 +0000
  • 934d4b93a2 VariantContext to VCF converter. BeagleROD, and phasing of VCF calls. Integration tests galore :-) depristo 2010-02-09 19:02:25 +0000
  • 369cc50802 Added playground walker that does a basic concordance check between two VCF files - an eval and a truth file - across all samples in the eval file. Produces per-sample, per-locus debug info and simple concordance stats. This is not meant to be extended, but rather used for validating the HapMap to VCF conversion in preparation for retiring GFF-based HapMap data. andrewk 2010-02-09 02:41:18 +0000
  • 94f892ad42 VCF->beagle and VCF phasing using beagle input. Appears to work fairly well. VariantContexts now support phased genotypes. depristo 2010-02-09 01:22:05 +0000
  • 457568485a simple Beagle input ROD depristo 2010-02-09 01:21:04 +0000
  • 57b8c9a53c Supporting infrastructure for merging SAM files. Not yet integrated into the datasource. hanna 2010-02-08 23:59:38 +0000
  • fc810a1800 Updated VCF Reader to parse VCFs according to the VCFv3.3 spec. Column headers are tab separated since sample names might have spaces. Updated test files in /humgen/gsa-scr1/GATK_Data/Validation_Data/*.vcf to remove spaces except for when they are supposed to be in the sample name. Added @Test before VCFReaderTest.testHeaderNoRecords() kshakir 2010-02-08 22:55:59 +0000
  • 935e76daa1 Minor changes to oneoff walkers. PlinkRod altered but still commented. chartl 2010-02-08 18:49:56 +0000
  • 21369869b7 Extend regex that supports every 'word' character to use any printable character except ':'. hanna 2010-02-08 03:29:55 +0000
  • 4fe851a83d Optimization: don't keep scoring an alternate consensus if it's already worse than the best alt seen so far. ebanks 2010-02-07 05:06:32 +0000
  • ca1917507f Various improvements and fixes: ebanks 2010-02-07 04:21:04 +0000
  • 57a168c0db Added a header crediting the python script as a source. Looping over an arbitrary number of headers. kshakir 2010-02-07 03:38:30 +0000
  • 3b1ab86d11 Added generic interfaces to RefMetaDataTracker to obtain VariantContext objects. More docs. Integration tests for VariantContexts using dbSNP and VCF. At this stage if you use dbSNP or VCF files only in your walkers, please move them over to the VariantContext, it's just nicer. If you've got RODs that implemented the old variation/genotype interfaces, and you want them to work in new walkers, please add an adaptor to VariantContextAdaptors in refdata package. It should be easy and will reduce burden in the long term when those interfaces are retired. depristo 2010-02-06 16:26:06 +0000
  • 995d55da81 now uses the new RMDT getVariantContext() functions instead of doing the work itself. depristo 2010-02-06 16:23:06 +0000
  • 33760834d6 commented out inactive (due to string ==) but actually incorrect code. Sometimes two wrongs do make a right depristo 2010-02-06 16:22:26 +0000
  • c7e006a996 Bug fixes for interval batching in sharding system. Sharding system now batches intervals and passes basic tests for small and large intervals and intervals that cross bin boundaries. Currently works only with a single BAM file. hanna 2010-02-05 21:47:54 +0000
  • fdc2789322 Minor fix: removed use command that doesn't work in /bin/sh scripts. andrewk 2010-02-05 19:52:50 +0000
  • 61a67cdce4 Moving file up a directory for dependencies andrewk 2010-02-05 19:36:13 +0000
  • 58456822ab Two perl scripts (from Kristian Cibulskis) and one python script for annotating VCF files with the information generated by the cancer MAF annotation tool. andrewk 2010-02-05 19:25:46 +0000
  • a1d5a384f4 Reverting the last reversal. bestConsensus points to something also kept in a set, so just reassigning it will NOT automatically destroy the underlying data; explicit clearing of unneeded data reinstated. STUPIDO!!! asivache 2010-02-05 18:08:53 +0000
  • cf7e6d0c0b Memory-saving change, same as in old IntervalCleaner (if alt consensus does not beat the best one, destroy its data immediately) asivache 2010-02-05 18:05:04 +0000
  • df0be25afb ooops, no need to destroy old best's data explicitly, it will be done automatically of course asivache 2010-02-05 18:03:16 +0000
  • 9f44018b7d Reducing memory footprint: if alt consensus does not beat the best alt observed so far, destroy its data immediately, instead of keeping them around. If new alt is better than the old best, then destroy the old best right away instead. asivache 2010-02-05 17:58:54 +0000
  • be33d1852c Reverting rpoplin 2010-02-05 15:57:09 +0000
  • af8c47fc2f Fixing up testVariantContext for integration tests for variant context. Printing of VCs and genotypes now stable using sorting. Cleaned up comments in quality score by strand. RefMetaDataTracker now directly allows walkers to obtain VariantContexts using the simple Collection<VariantContext> getAllVariantContexts(GenomeLoc curLocation, EnumSet<VariantContext.Type> allowedTypes, boolean requireStartHere, boolean takeFirstOnly) function. VCF and dbSNP VariantContexts now officially supported. Other importan types can be added to the adapator system in refdata package. Integration tests later today depristo 2010-02-05 15:42:54 +0000
  • 0d8d6e0a14 Ti/Tv module in VariantEval shows known and novel ratios if possible rpoplin 2010-02-05 15:37:40 +0000
  • 1494dc875f fixing up tests. Moves are complete depristo 2010-02-05 14:24:00 +0000
  • c6d86da4b8 almost managed to move things around perfectly in move go depristo 2010-02-05 14:18:26 +0000
  • 9e5c18d50f paranoid creation of directory before file move depristo 2010-02-05 14:04:16 +0000
  • e0af3bf761 updating back names depristo 2010-02-05 13:53:45 +0000
  • 777617b6c7 managed to actually move the files too! Damn you svn depristo 2010-02-05 13:47:19 +0000
  • 8938a4146d moving varianteval2 to it's own dir depristo 2010-02-05 13:37:04 +0000
  • 69132c81aa Documentation. Plus nicer structure to adaptors. Intermediate checkin before move into core depristo 2010-02-05 13:33:27 +0000
  • e53432d54d Checkpoint for combining adjacent intervals into the same shard. hanna 2010-02-05 02:48:02 +0000
  • 0d347d662a More plumbing: if after the shift window contains indel(s) at the first position, do not throw an exception, just print the warning (we can not deal with this situation!!) and discard those indels without trying to call them. This situation will most probably arise after forced shift over a messy region anyway. asivache 2010-02-04 21:06:28 +0000
  • 1d86dd7fd1 Interface changes following Matt's advice. VariantContexts are now immutable, and there are special mutable versions, in case you need to change things. AttributedObject now a InferredGeneticContext and package protected. VariantContexts are now named, which makes them easier to use with the rod system depristo 2010-02-04 20:55:49 +0000
  • e7b710791f OK, we finally ran into a messy dataset where we can not find a place to shift the window to: there's an indel at every position. Don't panick, don't throw an exception, just ignore the whole window completely, we do not want to call there. asivache 2010-02-04 19:49:56 +0000
  • 152f65b362 Do not die in --cycleOnly mode when the lane is not paired end, just count all single end basequals into the first column and leave the second column filled with 0s asivache 2010-02-04 19:48:12 +0000
  • a3cd56897d moving older versions of the oneoff project to archive, bye-bye asivache 2010-02-04 19:46:27 +0000
  • f7e7bcd2ef Oneoff project, totally unrelated to anything asivache 2010-02-04 19:44:50 +0000
  • 334da80e8b Fixed Mark's bad checkin. hanna 2010-02-04 12:40:58 +0000
  • 1ce0f06216 temp checkin for reorganization depristo 2010-02-04 11:10:24 +0000
  • 83b9d63d59 1. Added functionality to the data sources to allow engine to get mapping from input files to (merged) read group ids from those files. 2. Used said mapping to implement N-way-in,N-way-out functionality in the new indel cleaner. Still needs more testing (to be done after vacation but preliminary tests look good). 3. Fixes to VCF validator: ignore case when testing VCF reference base against true reference base and allow quals of -1 (as per spec). ebanks 2010-02-04 04:12:49 +0000
  • 233a652161 Making the dotted quartile lines more clear. rpoplin 2010-02-03 22:23:09 +0000
  • 210c4c9913 AnalyzeAnnotations now makes plots for the value in the QUAL column as if it were an annotation. rpoplin 2010-02-03 20:33:15 +0000
  • 280f38f635 Bringing IndelGenotyper back to life asivache 2010-02-03 20:12:49 +0000
  • 3f35e181d5 Add an alternate implementation of the BAM file reader that keeps the entire index in memory. Initial revision of BAMFileStat, a tool to inspect BAM file BGZF blocks and index entries. hanna 2010-02-03 19:48:15 +0000
  • c89ba7b1a4 improvements to variant eval 2. Now has titv calculations and mendelian violation detect support. we only make ~80 mendelian violations in 380K calls for the YRI trio, in case you are interested depristo 2010-02-03 16:03:19 +0000
  • af7cd9cf58 some very old tests relied on cancer data that got moved. Reset one to use data in the validation directory, the other to the artificial sam utils (the best approach). aaron 2010-02-02 23:13:10 +0000
  • fa2cd432fd better printing in VE2. Added support for TiTv analysis depristo 2010-02-02 21:20:29 +0000
  • cbbc0e98d2 fix for broken imports depristo 2010-02-02 15:20:27 +0000
  • 681c196097 V2 of VariantEval2. Framework is essentially complete., very simple and clear now compared to VE1. Support for any number of JEXL expressions. dbSNP% evaluation added to show paired comparison evaluation. Pretty printing output tables. Performance is poor but can easily be fixed (see todo notes). depristo 2010-02-02 14:18:46 +0000
  • ed23e54abd Added VariantEval to the GATK release. Woo hoo. hanna 2010-02-02 02:38:39 +0000
  • 9dbdfff786 Moved VariantEval to core. Updated integration test md5s to reflect new Analysis class names. hanna 2010-02-02 00:22:15 +0000
  • 4ddbaeed07 In attempt to reuse: --pairCountsOutput is now optional, if not specified then only per-locus statistics is collected; --silent - do not echo results into stdout; --minMapQ - count only bases coming from reads mapped with specified quality or better; --blacklistedlanes - do not count reads/bases coming from specific lanes. asivache 2010-02-01 22:05:19 +0000
  • 2c4f709f6f Bunch of oneoff stuff that I don't want to lose. Also: chartl 2010-02-01 21:35:10 +0000
  • 421282cfa3 Convenience method: getMappingFilteredPileup(int minMapQ) asivache 2010-02-01 21:19:53 +0000
  • 506d39f751 The UG calculations are now driven by an independent engine. This completely separates the genotyper walker from other walkers. ebanks 2010-02-01 20:57:31 +0000
  • d8e75cf631 Fix for Kiran's memory issue running UG...turned out to be a particularly bad interaction between @By(Reference) traversals and TreeReduce. hanna 2010-02-01 20:27:06 +0000
  • d9671dffba Documentation for VariantContext. Please read it and start using it. depristo 2010-02-01 17:49:51 +0000