b3c3023c3cAllows callers to handle HLA reference files as input (rather than hard-coded paths)
sjia
2010-06-14 18:56:08 +0000
266a47d83dThis file automaticially generates data and graphics for tearsheets and data processing reports
corin
2010-06-14 18:39:23 +0000
9666d47d17ooops, debug print now removed
asivache
2010-06-14 18:07:12 +0000
abdc8521eaAdded debug options for FindClosestHLAWalker
sjia
2010-06-14 17:52:03 +0000
c38390eabbAdded option for min number of matches between reads and alleles required to consider reads.
sjia
2010-06-14 16:08:49 +0000
4ab1f440c3A new argument: --targetIntervalsSorted (boolean flag). If specified, the interval file is assumed to be sorted (duh!) and it is NOT slurped into the memory but instead traversed directly on disk as needed. If the file turns out to be unsorted, an exception will be thrown at the point where inconsistency occurs (can be late into the processing!).
asivache
2010-06-14 16:00:22 +0000
671ac00748A simple utility class that implements a merging Iterator<GenomeLoc> built over an interval or bed file (this is NOT a rod, but rather a direct line-by-line file reader that converts strings to genome locs on the fly and merges overlapping intervals)
asivache
2010-06-14 15:54:37 +0000
f137bf8f85now adaptor silently skips empty lines in the underlying string iterator
asivache
2010-06-14 15:35:07 +0000
5704294f9dHLA caller updated - now searches all (common and rare) alleles, more efficient read filtering and allele comparison runs.
sjia
2010-06-14 15:14:40 +0000
7b7d3341f0trivial refactoring: isFile renamed to isIntervalFile and made public
asivache
2010-06-14 14:02:23 +0000
c3b68cc58dRethinking DownsamplingLocusIteratorByState with a flattened read structure. Samples are kept independent while processing, and only merged back in a priority queue if necessary in a special variant of the ReadBackedPileup. This code is not live yet except in the case of naive deduping. Downsampling by sample temporarily disabled, and the ReadBackedPileup variant is sketchy and not well integrated with StratifiedAlignmentContext or the walkers. Cleanup to follow.
hanna
2010-06-13 01:47:02 +0000
804facb0ccRemoving these utilities as part of a hostage negotation with Matt. Can I have my journal club paper now?!
kiran
2010-06-11 21:41:29 +0000
c0370f4d0aAdded both inclusive and exclusive filters
weisburd
2010-06-11 18:40:41 +0000
e6d8faf293making 'parseLocation' public static - as simple as the logic is, it's better kept in one place and I need it!
asivache
2010-06-11 18:19:59 +0000
8c28be5933Fixing a VCF bug for Sendu: we weren't emitting flags (booleans) correctly in VCF3.3 (rev'ed tribble for this). Updated dbsnp/hapmap membership info fields to be flags now instead of ints. While I was there, I added the change in the Annotator for Jan to force reads to be from a specific sample.
ebanks
2010-06-11 16:42:06 +0000
dde93e743falways output a brief test summary to the screen, and xml to disk
aaron
2010-06-11 16:24:50 +0000
22620ba95cAdding "abi_solid" to the list of known platforms.
ebanks
2010-06-11 13:37:19 +0000
63ad71cca6Fix busted code. Note for all: String.valueOf(byte[]) doesn't work. You must use new String(byte[]).
ebanks
2010-06-11 05:01:48 +0000
338bb9adf4CommandLineProgram for measuring java I/O speeds for large plain-text or gzipped files.
weisburd
2010-06-10 21:34:37 +0000
d1a4c4f0d3Added -w filter option allowing user to specify chromosomes to be skipped.
weisburd
2010-06-10 20:58:25 +0000
06fc5eecf8Implemented TreeReducible - if num threads > 1, the output will be accumulated in memory and written to a vcf file at the end - in onTraveralDone(..). If num threads == 1, things will work as before - where vcf records are written to disk as soon as they are computed with map(..).
weisburd
2010-06-10 20:57:23 +0000
3b375cb237Sped up parseGenomeLoc(..) by replacing regexp with String.indexOf(..) - attempt 2
weisburd
2010-06-10 20:54:36 +0000
e27951ab39re-updating the VCF code to handle spaces in sample names
aaron
2010-06-10 20:18:34 +0000
99b684ea89Adding new support for reference data. ReferenceDataSource is a new class that manages reference data, and allows IndexedFastaSequenceFile to be a simple reader. This checkin also includes FastaSequenceIndexBuilder, which reads a fasta file and creates an index, like samtools faidx. Right now this is not enabled, because we are still working out thread safety. So the only new UI change is that GATK can be run without a fai file. Soon, we will enable 1) GATK to be run without a dict file too, and 2) both dict and fai files will be saved on disk for future program executions. For more info, see ReferenceDataSource.java
bthomas
2010-06-10 20:10:23 +0000
f55f32d4eeBug fix.
hanna
2010-06-10 01:53:26 +0000
ca4eab1d23Now annotations that require reads return null if there's no alignment context, so that running without reads adds annotations only for the appropriate fields. Added an integration test for the read-less case.
ebanks
2010-06-09 20:36:46 +0000
6941c81bfareverting revision 3522 to the old code until we fix the tests.
aaron
2010-06-09 19:25:02 +0000
dbee21a50fBugfixes for the case when no read groups / no samples are available.
hanna
2010-06-09 18:47:05 +0000
adc4c4e577Sped up parseGenomeLoc(..) by replacing regexp with String.indexOf(..)
weisburd
2010-06-09 18:11:43 +0000
20167fd411Final changes to MVC -- associates variants with regions of homozygosity in child and parents, corrects for genotype errors, and prints out a separate file with informationf or each region of homozygosity.
chartl
2010-06-09 18:05:37 +0000
4f00e265a8quick update for a change I implemented for Ryan
aaron
2010-06-09 17:23:31 +0000
ad98512f6cadding changes so that we look at the headers already loaded by the engine for samples and other VCF utils, and not create readers for each file to get them (this caused Tribble to regerenate indices if the index file can't be written to disk).
aaron
2010-06-09 17:21:12 +0000
c1b7bcc786Fixed handling of mitochondrial genes - added special cases such as ATT being a start codon in mitochondria. Added warning if a gene doesn't start with Met or end in a stop codon
weisburd
2010-06-09 17:15:47 +0000
6fd2d39a7dModified run_locally mode to use os.system(..) instead of popen
weisburd
2010-06-09 17:10:03 +0000
a3ccf49f5bWrite error to stderr
weisburd
2010-06-09 17:09:10 +0000
9b2fcc4711Refactoring of the annotation system: 1. VA is now a ROD walker so it no longer requires reads (needs a little more testing) 2. Annotations can now represent multiple INFO fields (i.e. sets of key/value pairs) 3. The chromosome count annotations have been pulled out of UG and the VCF writer code and into VA where they belong. Fixed the headers too.
ebanks
2010-06-09 17:05:51 +0000
84563b37e5Partial flattening of the hanger data structure. Hanger data structure is not currently as flat as it could / should be, but it's already comparable to the speed of the reference implementation.
hanna
2010-06-09 16:28:49 +0000
8f9e3e8ad7Commit for Kiran; but this is now working, barring little exceptions that I've yet to run across...
chartl
2010-06-09 14:21:19 +0000
6febd0291drev tribble to include some dbsnp clean-up and fixes
aaron
2010-06-09 03:08:31 +0000
6d5556939dupdating Tribble with a couple of important Tabix fixes, and updating the variant eval integration tests to run each test with both plain vcf and gzipped tabix (added the tabix version to the vlidation directory), using the same md5sum.
aaron
2010-06-09 01:47:04 +0000
2b31975cb4Added more options for coordinate systems - now you can add 1 to either the start coordinates, the end coordinates, or both
weisburd
2010-06-08 22:49:19 +0000
410afcdf2cAdded parallelization options - when running locally, multiple processes can be spawned, or a -nt arg can be specified to run each TranscriptToInfo instance multi-threaded
weisburd
2010-06-08 22:48:07 +0000
92c72d3361Added back lines that update the *big-table-header.txt file before using it
weisburd
2010-06-08 22:45:41 +0000
3c24223d02Script for concatenating 2 AnnotatorInputTables, and writing the result to standard out. Merge-sorts the 2 tables while concatenating them
weisburd
2010-06-08 22:44:16 +0000
c2858c8988Minor performance enhancement. Checkpoint commit before major performance overhaul.
hanna
2010-06-08 21:39:39 +0000
5ed2818ffbForgot to commit code i relied upon
chartl
2010-06-08 21:01:35 +0000
736098b58dA quick commit before running home. This is a re-factored version of the OppositeHomozygoteClassifier which will work with deNovo violations as well. Some code still needs to be migrated from OHC which is wy that walker isn't yet deleted. This'll be up and running tonight.
chartl
2010-06-08 20:47:01 +0000
de134c226dRemoved ability of users to specify annotations to recompute, cleanups.
delangel
2010-06-08 19:17:59 +0000
4d1a6b3d99quick changes for G
ebanks
2010-06-08 16:33:27 +0000
907931c902a) Update annotations when creating new vcf with Beagle's imputed data. Since genotypes may (will) change based on imputation, several annotations need to be updated. By default, AC, AF, AN and AB will be updated. User can force extra annotaqtions to be updated with -A <annotation> argument.
delangel
2010-06-08 15:12:04 +0000
933133ee28Initial commit of the opposite homozygote classifier. Currently does the following, given a trio vcf: + Identifies opposite homozygote sites + Identifies the parent from whom it is expected that a null allele was inherited (or whether it was a putative genotype error; e.g. mom=homref, dad=homref, child=homvar) + Labels each opposite homozygote with its homozygous region in the child (e.g. region 1, region 2) + Labels each opposite homozygote with the size of the homozygous region in which it was found, the number of child homozygotes in the region, and the number of opposite homozygote violations within that region
chartl
2010-06-08 03:56:07 +0000
199e4208cdBug fixes.
hanna
2010-06-08 00:30:33 +0000
52ab9f2417Feature parity between LocusIteratorByState, DownsamplingLocusIteratorByState, including pushing mrl / the LocusOverflowTracker into LocusIteratorByState. Note that the 'Matt Hanna exception', is still enabled because I haven't yet validated the performance of the DownsamplingLocusIteratorByState when running without downsampling.
hanna
2010-06-07 22:58:21 +0000
5c4d070566Push Mark's changes in LocusIteratorByState into DownsamplingLocusIteratorByState in preparation for merging the two into one.
hanna
2010-06-07 17:29:30 +0000
6eeb1693caJEXL2 upgrade. Improvements to JEXL processing including dynamically resolving variable -> value bindings instead of up front adding them to a map. Performance improvements and code cleanup throughout.
depristo
2010-06-07 00:33:02 +0000
c1ecf75dd5Update to the latest rev of the picard sharding patch. Includes updates reflecting the imminent move of IlluminaUtil into picard public.
hanna
2010-06-06 20:33:21 +0000
d4c66d6191a) Small cleanup b) Fix major issue with Beagle likelihood converter: if likelihood triplets from UG end up being too low, then Beagle input file will be produced with 0.00,0.00,0.00 triplet. If all samples at a marker have this issue, Beagle will effectively produce junk. To fix, likelihoods are renormalized before converting to linear space.
delangel
2010-06-06 17:31:59 +0000
cfa18f6743Fixing missed update with new Allele in it
depristo
2010-06-04 23:56:34 +0000
3ea506fe52No more new Allele() -- must use create. Allelel simple alleles are now cached for efficiency reasons. VCF4 codec optimizations -- 4x performance in general. Now working in general but hooked up to the ROD system now as VCF4. WARNING -- does not actually work with indels, genotype filters, etc.
depristo
2010-06-04 23:03:55 +0000
ef47a69c50a) First fully functional (sort of) version of walker that parses Beagle imputation output files and produce a vcf with imputed genotypes. More doc/info to follow shortly. Issues still to be solved: a) Walker changes all genotypes based on Beagle data, but annotations on the original VCF are unchanged. They should in theory be recomputed based on new genotypes. b) Current implementation is ugly, dirty unwieldy and will necessitate a refactoring soon so I can keep my pride. Most aesthetically affronting issue right now is that we read the full Beagle files at initialization and keep them in memory, but a more delicate implementation would just read from files on a marker by marker basis. Issue that currently prevents this is that BufferedReader() instances don't seem to play nice when called from the map() function.
delangel
2010-06-04 20:37:25 +0000
b811e61ae1Optimized, nearly complete VCF4 reader 2-4x faster than the previous implementation, along with a VCF4 reader performance testing walker that can read 3/4 files, useful for benchmarking
depristo
2010-06-04 18:11:38 +0000
6482b87741adding the super experimental, half-broken, generally crippled, awkwardly commented, header ignoring vcf4 code. Don't use this, unless you're a developer for VCF4. If so, remove the exception from the constructor so that it won't always exception out.
aaron
2010-06-04 07:38:46 +0000
0b03e28b60updating the tribble library to include the reference dictionary reading / writing. We now check the dictionaries of any tracks that have them against the reference (all new tribble tracks and out-of-date tracks will have this). Also renamed some classes to be more reflective of their function.
aaron
2010-06-04 06:34:26 +0000
3d055e3d16Fail fast if users try to parallelize a read walker.
hanna
2010-06-03 18:14:33 +0000
7d79848f40Better error message when bam file / list file with wrong extension is supplied.
hanna
2010-06-03 17:52:48 +0000
597b3744abAlways use phasing info when converting genotypes to strings
ebanks
2010-06-03 17:50:50 +0000
e2b41082afGATK now does automatic adaptor filtering in locus iterators (but not expt. downsampling iterator). General support for LocusIteratorFilters just like read filters but only applying at particular bases. Updated tools with new MD5 sums due to adaptor bases in their integrationtest data. Not that as a side effect here reads close to each other with odd orientations are also filtered out. Updated minor argument to VariantRecalibrator to change the qStep value on the command line
depristo
2010-06-02 22:26:32 +0000
8ec091d6d2re-enabling regeneration of the tribble index if it's out of date. Also moved the class that can detect text in the log4j stream (useful in testing to make sure appropriate messages are generated).
aaron
2010-06-02 17:45:51 +0000
f0c379dde8Unconsequential changes in report formatting
asivache
2010-06-02 17:43:25 +0000
3ab936181cSupports the join feature of GenomicAnnotator
weisburd
2010-06-02 16:29:57 +0000
12c3e3ecdaAdded back the check for values.size() != header.size(). Now exception will be thrown if number of columns in a record doesn't equal number of columns in the header
weisburd
2010-06-02 16:23:05 +0000
290771a8c2Automatic cutting of recalibrated variant calls using ApplyVariantCuts. VariantRecalibrator produces the tranches plot alongside the optimization curve. Specify the levels using -tranche 1.0 -tranche 5.0 etc
rpoplin
2010-06-02 15:03:00 +0000
4a555827aaRemoving more toUpperCase sanity checks
ebanks
2010-06-02 14:38:39 +0000
56e504789atrivial change: toUpperCase no longer necessary
ebanks
2010-06-02 14:00:47 +0000
87fe60fe4fFix for Sendu. new Process and p.waitFor() don't seem to work on his farm. Throws an IOException. This was a problem way back with AnalyzeCovariates too.
rpoplin
2010-06-02 11:37:10 +0000
7f0c638653Fix for the indel cleaner: I forgot to "unclip" the cigar string (even though the clipped bases were removed) before using it as an alternate consensus in a particular instance.
ebanks
2010-06-01 02:07:20 +0000
21427211c0Personal MD5 database system now live. WalkerTest now maintains a database of result files associated with MD5 results in integrationtest/, and provides command lines for diff-ing expected to current md5 results when encountering failed intergration tests. The suite currently takes 200Mb to store. Update and run intergrationtest to build your very own expectation database for future development work.
depristo
2010-05-31 16:06:16 +0000
2b02324587Support for detecting and automatically excluding reads reading into the adaptor sequence and, if desired, also only showing the first pair when two reads overlap in the fragment. Not enabled, an intermediate check in before updating and verifying the impact on locus walkers everywhere.
depristo
2010-05-30 18:00:12 +0000
e96bcf2128GSA-333 Replaced all accidental usages of getCanonicalPath (which should never be used) with getAbsolutePath.
kshakir
2010-05-28 22:27:05 +0000
eb25e41111minor update to new tribble name
ebanks
2010-05-28 20:23:25 +0000
ffeb3fd80dThanks to Guillermo, I found a bug in the Unified Genotyper output: GL was posteriors instead of likelihoods. Not a huge deal because the priors were flat, but fixed nonetheless. Also, needed to update Tribble. Minor updates to the Beagle input maker.
ebanks
2010-05-28 19:28:26 +0000
4e268ef6acRemoving the Variant Recalibration Performance test because it isn't ready yet.
rpoplin
2010-05-28 18:27:25 +0000
522dd7a5b2Adding the variantrecalibration classes.
rpoplin
2010-05-28 18:21:27 +0000
2014837f8aVariantOptimizer package is moved to core, renamed as VariantRecalibration, and added to the binary release package. VariantOptimizer walker is renamed to GenerateVariantClustersWalker and ApplyVariantClustersWalker renamed to VariantRecalibrator. Integration tests added, performance tests still to be done.
rpoplin
2010-05-28 18:20:18 +0000
871cf0f4f6Call out ROD types by there record type, instead of the codec type (which was clumsy). So instead of:
aaron
2010-05-28 14:52:44 +0000
hanna
sjia
weisburd
corin
asivache
kiran
ebanks
aaron
bthomas
chartl
delangel
depristo
rpoplin
kshakir