df4e001a07Renamed to more accurately describe its function.
ebanks
2009-12-03 17:34:49 +0000
c2017cc91bPrintCoverageWalker functionality moved to DepthOfCoverageWalker. Added integration tests.
ebanks
2009-12-03 17:23:59 +0000
608fa4cc3aUpdate packaging system to produce multiple jars if asked nicely.
hanna
2009-12-03 17:11:12 +0000
01cf5cc7411. Merged CoverageHistogram into DepthOfCoverageWalker 2. Fixed bug in histogram calculation for small intervals 3. Better output in DoCWalker 4. Comments added to code
ebanks
2009-12-03 17:01:53 +0000
44b9f60735PercentOfBasesCovered functionality moved to DepthOfCoverageWalker. Added integration tests.
ebanks
2009-12-03 16:11:09 +0000
126d1eca35Move to core (qc/)
ebanks
2009-12-03 15:45:58 +0000
9da5cc25adMore archiving (with permission from Andrey) plus a move to core.
ebanks
2009-12-03 15:40:27 +0000
b3bdcd0e60make sure we close the error log stream in CommandLineProgram if it's opened; unit tests and clean-up for BasicVariation
aaron
2009-12-03 06:59:27 +0000
d27d56ea0cadding the option to not halt after a test failure; it will still report failure at the end. This makes sense for Bamboo so we know all the tests that fail, not just the first. It is also useful for checking what tests are affected by a change.
aaron
2009-12-03 06:19:39 +0000
a88202c3f6Refactored DoCWalker to output in a more helpful and usable style. It now outputs in tabular format with 2 different sections: per locus and then per interval. I am now at a point where I can merge the functionality from other coverage walkers into this one. Thanks to Andrew for input.
ebanks
2009-12-03 05:28:21 +0000
d7e4cd4c82Moving some useful and stable walkers to core: - ClipReads - PrintRODs (generalized to print all RODs that are Variations) - FixBAMSortOrderTag (added documentation to walker so that people know what it does and why)
ebanks
2009-12-03 03:00:45 +0000
2460f2a789minor update to build.xml for CG utilities. Hopefully bamboo will be back online
depristo
2009-12-03 00:06:02 +0000
0b6b92bf5bFix (hopefully) the rest of the problems with the build.xml svn merge.
hanna
2009-12-02 23:53:31 +0000
f41965a73eFixed incompatiblity in variable name between my script and the new elements merged into my script.
hanna
2009-12-02 23:40:50 +0000
357abc9e89Update build.xml to make it oneoffprojects-aware, and do some misc formatting cleanup and moves. Here goes nothing.
hanna
2009-12-02 23:27:45 +0000
46f3d3e39bAdded comments to AnalyzeCovariates and R scripts. R script prevents residuals from going off the edge of the plot. Added skeleton code to the recalibration walkers showing how we plan to handle SOLID reference inserting behavior.
rpoplin
2009-12-02 23:15:52 +0000
451a20ed55commenting out some broken integration tests, to be uncommented if needed.
aaron
2009-12-02 23:13:24 +0000
a3343c75dbMove and rename a hybrid-selection-specific coverage calculation to hybridselection/
ebanks
2009-12-02 20:11:22 +0000
2c83f2f2bcMove MSG - plus now obsolete classes which it depends on -- to oneoffprojects (with permission from Jared).
ebanks
2009-12-02 20:04:22 +0000
9c597309d3Changed the sizes of the dots and bars on the plots generated by the R script which is called from AnalyzeCovariates.java
rpoplin
2009-12-02 19:47:48 +0000
93842abb4aLet the packages know about the changes made for AnalyzeCovariates.jar
rpoplin
2009-12-02 19:09:53 +0000
c180a76b05Added option "append": if set, and the specified discovery output already exists, don't re-call anything that's already present in that file. Append new calls to it.
jmaguire
2009-12-02 18:56:19 +0000
0a2304eff8- Rename minConfidenceScore in VariantEval to minPhredConfidenceScore - Moved validation walkers to new qc dir - Killed unused test
ebanks
2009-12-02 17:59:19 +0000
a5dfc9107d- Cleaned up annotation code some more - Use QualityUtils when phred-scaling now
ebanks
2009-12-02 17:45:29 +0000
7055a3ea2d- All annotations are now required to return their VCF INFO keys and descriptions - Renamed keys to fit with the standard naming - FisherStrand is no longer standard - Integration tests no longer test experimental annotations since they're not stable
ebanks
2009-12-02 17:24:06 +0000
67179e2412Initial checkin of AnalyzeCovariates.java which replaces analyzeRecalQuals_1KG.py and is updated to use the new Covariates system. It creates similar plots of residual error for each covariate that was used in the calculation. There is also an option to filter out base qualities below a given threshold.
rpoplin
2009-12-02 16:47:35 +0000
2838629724-VCF writer now checks whether the allele frequency has been set before trying to write it out. -Renamed methods to be more consistent.
ebanks
2009-12-02 16:25:32 +0000
6231637615fixes for VariantAnnotations and second bases. Misc. removal of failing (and unstable) integration tests that require rereview
depristo
2009-12-02 15:41:35 +0000
886c44303a-Removing BTTJ integration test -- this broke a few revisions ago (2169) and it is unclear whether the resulting change was a correction to something that had previously been incorrect, or a true build-breaker. I'm currently investigating which case this is, but since Bamboo is back up I'm removing this _temporarily_ so that other testing can occur, and will make whatever changes to the test necessary to reflect the truth, then replace the test itself. Additional (and related) pileup tests are upcoming as well.
chartl
2009-12-02 05:37:15 +0000
b979bd2ced- Optimized implementation of -byReadGroup in DoCWalker - Added implementation of -bySample in DoCWalker - Removed CoverageBySample and added a watered down version to the examples directory
ebanks
2009-12-02 03:39:24 +0000
7c73496e72Moved DoC walker over to new pileup system so it no longer moves like it's stuck in molasses.
ebanks
2009-12-02 02:46:39 +0000
ba8a8febc6Thanks to Steve Hershman for finding this bug: getNegLog10PError() does not equal the confidence score (you need to multiply by 10 as confidence is traditionally phred scaled). Probably we should change the method to be getNeg10Log10PError(). Anyone have strong feelings on this?
ebanks
2009-12-02 01:59:03 +0000
bc8f3d58aeAdded an aligner package to make it easier for people to use outside the GATK.
hanna
2009-12-01 23:04:10 +0000
3303808a8fYet more walkers moved to oneoffprojects. Made hybridselection subdir in playground.
ebanks
2009-12-01 21:29:12 +0000
05923f7fbaStarted transition to oneoffprojects. Moved/killed a few other walkers (with permission).
ebanks
2009-12-01 21:19:02 +0000
c36069355eTrivial change to verbose
ebanks
2009-12-01 20:48:10 +0000
74f6526e09VCFHomogenizer: A class that extends InputStream and dynamically re-writes pilot1 VCF's to be on-spec.
jmaguire
2009-12-01 17:55:42 +0000
adf8f1f8b3Add an InputStream constructor, which is immensely useful for various reasons.
jmaguire
2009-12-01 17:25:00 +0000
e581cceab6Got Kris's permission to delete these walkers.
ebanks
2009-12-01 16:57:28 +0000
3180fffd43Eliminated unnecessary boxing of longs in RecalDatum. Changes to RecalDatum in preparation for new AnalyzeCovariates script. Updated TableRecalibrationWalker to make use of these changes.
rpoplin
2009-12-01 16:49:05 +0000
c93d37d9fbcontinuing improvements in output of snpSelector
depristo
2009-12-01 15:42:06 +0000
21a9a717e4Some minor changes and test:
chartl
2009-12-01 15:15:54 +0000
084337087eRemoving deprecated code and walkers for which I had the green light from repository. Moved piecemealannotator and secondarybases to archive.
ebanks
2009-12-01 05:58:20 +0000
2c16c18a04Move Andrey's old indel code (plus MSG accuracy test, which depends on it) to archive.
ebanks
2009-12-01 05:29:00 +0000
7c6c490652An unfinished implementation of the Wilcoxon rank sum test and a variant annotation that uses it. I need to merge and update this code with Tim's implementation somehow - but that won't happen until later this week, so I'm committing this before I accidentally blow it away.
ebanks
2009-12-01 04:56:17 +0000
00f15ea909Improved performance of deletion-free pileup and added mapping-quality-zero-free pileup convenience method. Finished converting genotyper and annotator code to new ReadBackedPileup system.
ebanks
2009-12-01 04:50:47 +0000
4969cb1957CountCovariates uses new optimized ReadBackedPileup. It also smarter about re-doing calculations for the dnsnp variation rate sanity check.
rpoplin
2009-11-30 20:35:40 +0000
add2fa7ab4more use of new ReadBackedPileup optimizations
ebanks
2009-11-30 20:04:01 +0000
f5fe28cc28Another matlab script -- this time for making power and coverage plots over a specific gene region. Lots of fun file reading, string manipulation, and exploration of the set() function
chartl
2009-11-30 20:02:25 +0000
817e2cb8c5Recalibrator makes use of the new GATKSAMRecord wrapper and now no longer has to hash the SAMRecord. Covariate's getValue method signature has changed to take the SAMRecord instead of the ReadHashDatum. ReadHashDatum removed completely.
rpoplin
2009-11-30 19:59:17 +0000
e9a8156cfbUse new optimized ReadBackedPileup
ebanks
2009-11-30 18:17:18 +0000
d8146ab23dChanged the format of the recalibration csv file slightly so that it is easier to load the file into something like R and look at the values of the covariates.
rpoplin
2009-11-30 17:55:23 +0000
a184d28ce9Completing the optimization started by Matt: we now wrap SAMRecords and SAMReadGroupRecords with our own versions which cache oft-used variables (e.g. platform, readString, strand flag). All walkers automagically get this speedup since the wrapping occurs in the engine.
ebanks
2009-11-30 17:39:29 +0000
d5cda76c8bA new subversion repository for matlab scripts, so they don't go away in case of catastrophic Yang failure.
chartl
2009-11-30 16:25:07 +0000
af22ca1b47Bug fixes for VariantEval. dbCoverage now reports dbSNP rate, not some wierd eval_snps_in_db as before. We now separate non-indel and non-snp db sites in dbcoverage. Some dbSNP records don't fit into these two categories. Also fixed a consistency issue where novel / known sites where being determined solely by whether dbSNP had a record there, rather than the stricter dbcoverage screen for isSNP().
depristo
2009-11-30 01:39:01 +0000
2ea93385beBetter support for comparison to truth. Now emits FP rates for each covariate if a truth file is provided. Also now writes out a detailed recal.log file that can be parsed directly into R
depristo
2009-11-29 22:20:40 +0000
662bbbd53bAwful stupid bug. This will use up one of my bad code offsets.
chartl
2009-11-29 20:09:33 +0000
fa2d564f2cAnd the compulsory one-second-later fix -- better handling of arguments (e.g. for callng from outside of /trunk/python/)
chartl
2009-11-29 20:02:43 +0000
45673d7851A quick and dirty script that, given a list of input VCF files, will output a new VCF file which looks identical to the first VCF file of the input list, except that the info field has been updated to reflect the union of all the INFO annotations across the VCF files
chartl
2009-11-29 20:01:29 +0000
27651d8dc2Oops. numReads is now called size
chartl
2009-11-29 06:59:17 +0000
21744e024bQuick walker that determines % of bases covered at (user - defined depth)x . I've been maintaining it in my directories alone, but now that i've accidentally deleted it twice, into playground it goes
chartl
2009-11-29 06:51:19 +0000
3300ca906aAn iterator for Eric to use when injecting his new wrapping reads -- a stopgap solution for getting additional caching functionality into a SAMRecord.
hanna
2009-11-27 22:25:52 +0000
26db15be5cAdded SingleReadGroupFilter to only use reads from a specific read group, filtering out all others.
rpoplin
2009-11-27 20:33:59 +0000
d1298dda13Encapsulated the sections of code that were shared by the two Recalibration walkers. This includes both the shared command line arguments and the section of code in the map methods which pull out data from the SAMRecord and stuff it into the ReadHashDatum. Command line arguments are now passed to the Covariates using a new initialize method that all Covariates must implement. Updated the dbsnp sanity check warning message to be less cryptic.
rpoplin
2009-11-27 19:54:10 +0000
75b61a3663Updated, optimized REadBackedPileup. Updated test that was breaking the build -- it created a pileup from reads without bases...
depristo
2009-11-25 23:30:39 +0000
65da04ca85Now uses the theoretically correct relationship between SNP FP and TP ratios for Illumina data. maxQ score for a snp is now 60
depristo
2009-11-25 22:08:12 +0000
49f020b0b9Add to svn:ignore
alecw
2009-11-25 21:42:40 +0000
ac1b289d55Add tile to ReadHashDatum, and implement TileCovariate
alecw
2009-11-25 21:41:42 +0000
db40e28e54ReadBackedPileup in all its glory. Documented, aligned with the output of LocusIteratorByState, and caching common outputs for performance
depristo
2009-11-25 20:54:44 +0000
b44363d20aRemoved silly casts from Integer to int.
rpoplin
2009-11-25 19:59:21 +0000
84921b18edPush version number of picard-private-parts
alecw
2009-11-25 19:28:57 +0000
5f2801e015Push version number of picard-private-parts
alecw
2009-11-25 19:28:42 +0000
d563f4bd2cAdd IlluminaUtil to picard-private.jar
alecw
2009-11-25 19:25:21 +0000
d0f673f0c0Use Math.abs so we don't get (inconsistent) -0's
ebanks
2009-11-25 19:08:34 +0000
fe8809d12aAdded Covariate classes to the Early Access binary
rpoplin
2009-11-25 18:59:21 +0000
e8bb88e33dUpdated package for early access
ebanks
2009-11-25 18:08:12 +0000
6ff8526592Added arguments to the recalibration walkers so the user can specify the default read group id and platform to use when a read has no read group. There are also options to force every read group and every platform to be the specified values. Added integration tests that use a bam file with no read groups. Added comments to all the covariates to explain what each of the methods in the Covariate interface are used for.
rpoplin
2009-11-25 15:41:12 +0000
cfbd9332b0small cleanups for the GATK paper genotyper; switched to the managed output system.
aaron
2009-11-25 08:04:13 +0000
e1e5b35b19Don't have the spanning deletions argument be a hard cutoff, but instead be a percentage of the reads in the pileup. Default is now 5% of reads.
ebanks
2009-11-25 04:54:44 +0000
03342c1fddRestructuring and interface change to ReadBackedPileup. We now lower support the Pileup interface, the BasicPileup static methods, and the ReadBackedPileup class. Now everything is a ReadBackedPileup and all methods to manipulate pileups are off of it. Also provides the recommended iterable() interface of pileup elements so you can use the syntax for (PileupElement p : pileup) and access directly from p.getBase() and p.getQual() and p.getSecondBase(). Only a few straggler walkers use the old style interface -- but those walkers will be retired soon. Documentation coming in the AM. Please everyone use the new syntax, it's safer, and will be more efficient as soon as the LocusIteratorByState directly emits the ReadBackedPileup for the Alignment context, as opposed to the current interface. In the process of the change over, discovered several bugs in the second-best base code due to things getting out of sync, but these changes were resolved manually. All other integrationtests passed without modification.
depristo
2009-11-25 03:51:41 +0000
2cb3e53b0bVerbose mode shouldn't be printing out 'NaN's and 'Infinity's
ebanks
2009-11-24 22:01:00 +0000
c9ff5f209cAdded a CountCovariates integration test that uses a vcf file as the list of variant sites to skip over instead of the usual dbSNP rod.
rpoplin
2009-11-24 21:51:38 +0000
3484f652e71. Variation is now passed to VariantAnnotator along with the List of Genotypes so non-genotype calls has access to all relevant info. 2. Killed OnOffGenoype 3. SpanningDeletions is now SpanningDeletionFraction
ebanks
2009-11-24 21:47:20 +0000
e05cb346f3GenotypeLocusData now extends Variation. Also, Variations should be INSERTIONs or DELETIONs (and not just INDELs). Technically, VCF records can be indels now. More changes coming
ebanks
2009-11-24 21:07:55 +0000
dffa46b380BAM files created by TableRecalibration now have the version number and list of covariates used appended to their header with a new 'PG' tag. Eventually the entire list of command line args will be put in there as well. Big thanks to Matt and Aaron. The integration test uses the --no_pg_tag so that the md5 doesn't change every time the version number changes.
rpoplin
2009-11-24 20:53:57 +0000
ebanks
hanna
aaron
depristo
rpoplin
jmaguire
chartl
alecw