FYI: creating an exclusive lock on a file that does not exist will create that file as an empty file, and will NOT delete that file after the program terminates. So watch out if it's possible that the file you're locking does not exist - could end up leaving extra files that confuse users.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3795 348d0f76-0448-11de-a6fe-93d51630548a
2. Updated liftover code (and scripts) to emit vcf 4.0 and no longer depend on VCFRecord.
3. Beagle walker now also emits vcf 4.0.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3767 348d0f76-0448-11de-a6fe-93d51630548a
b) Updated VCF4WriterTestWalker to take either VCF3 or VCF4 as inputs (this walker can also be used to convert from 3.3 to 4.0).
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3711 348d0f76-0448-11de-a6fe-93d51630548a
- Fixed and cleaned code to produce trailing and padding bases in alleles around indels.
- Deal better with missing fields.
Pending:
- Chopping missing fields at end of genotypes.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3679 348d0f76-0448-11de-a6fe-93d51630548a
Base encoding for complex events is still brittle and most probably still has issues, fixes upcoming.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3671 348d0f76-0448-11de-a6fe-93d51630548a
Also, moved the flag indicating VCF4.0 to the VCFWriter constructor.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3669 348d0f76-0448-11de-a6fe-93d51630548a
(and loading time) of long reference sequences by better controlling the input buffer size.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3665 348d0f76-0448-11de-a6fe-93d51630548a
See VCF4WriterTestWalker for usage example: it just amounts to adding
vcfWriter.add(vc,ref.getBases()) in walker.
add() method in VCFWriter is polymorphic and can also take a VCFRecord, lthough eventually this should be obsolete.
addRecord is still supported so all backward compatibility is maintained.
Resulting VCF4.0 are still not perfect, so additional changes are in progress. Specifically:
a) INFO codes of length 0 (e.g. HM, DB) are not emitted correctly (they should emit just "HM" but now they emit "HM=1").
b) Genotype values that are specified as Integer in header are ignored in type and are printed out as Doubles.
Both issues should be corrected with better header parsing.
2) Check in ability of Beagle to mask an additional percentage of genotype likelihoods (0 by default), for testing purposes.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3664 348d0f76-0448-11de-a6fe-93d51630548a
Note: the -L argument takes both interval strings and filenames. If you specify an interval string that is also a file, an error will be thrown to move the file: ie. if you have a file "chr1" in the parent directory, GATK will ask you to move/delete it. But, this only happens with interval string arguments, NOT with intervals that are contained in files, which is a majority of the use case.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3602 348d0f76-0448-11de-a6fe-93d51630548a
Note that this functionality will only work if the directory with the fasta is writeable. GATK will fail if directory is read only and and either the .fasta.fai or .dict files don't exist. In the future, we could have these references be created in memory, but we decided against it this time.
Locking was also added to ReferenceDataSource so no issues come up while running multiple GATKs on the same reference: we don't want one process to be half-finished and another try to read it. So, you could see error messages related to locking. See ReferenceDataSource.java for explanation of the locking strategy.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3601 348d0f76-0448-11de-a6fe-93d51630548a
issue building up pileups from pileups of individual sample data.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3598 348d0f76-0448-11de-a6fe-93d51630548a
Provides a cleaner interface with extended events inheriting all of the basic RBP
functionality. Implementation is still slightly messy, but should allow users to
provide separate implementations of methods for sample split pileups and unsplit
pileups for efficiency's sake.
Methods not covered by unit/integration tests have not been sufficiently tested yet.
Unit tests will follow this week.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3597 348d0f76-0448-11de-a6fe-93d51630548a
are gone where I could identify them, but hierarchies that split to support two sharding systems have
not yet been taken apart.
@Eric: ~4k lines.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3580 348d0f76-0448-11de-a6fe-93d51630548a
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.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3540 348d0f76-0448-11de-a6fe-93d51630548a
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.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3513 348d0f76-0448-11de-a6fe-93d51630548a
@Requires(value={},referenceMetaData=@RMD(name="eval",type= VCFCodec.class))
you'd say:
@Requires(value={},referenceMetaData=@RMD(name="eval",type= VCFRecord.class))
Which is more in-line with what was done before. All instances in the existing codebase should be switched over.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3457 348d0f76-0448-11de-a6fe-93d51630548a
Small change for Ryan to get hard-clipped reads working for the recalibrator.
PLEASE DO NOT RELEASE THIS WEEK. I still have some more testing to do and need Mark to run WG jobs.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3430 348d0f76-0448-11de-a6fe-93d51630548a
as walker attributes or from the command-line. Not ready yet! Downsampling/deduping
works in a general sense, but this approach has not been completely optimized or validated.
Use with caution.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3392 348d0f76-0448-11de-a6fe-93d51630548a
the reference implementation. Also implemented a heap size monitor that can
be used to programmatically report the current heap size.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3367 348d0f76-0448-11de-a6fe-93d51630548a
which improved startup time 5-10x when dynamically merging many BAMs.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3331 348d0f76-0448-11de-a6fe-93d51630548a
VCFGenotypeWriterAdapter now explicitly uses the passed reference base instead of deriving it from VatriantContext (in SNP mode as well!), other writers simply ignore that additional argument.
SimpleIndelCalculationModel now WORKS (or rather, it does produce calls :) )
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3228 348d0f76-0448-11de-a6fe-93d51630548a
unicode quote characters embedded in it. These characters were invisible inside
IntelliJ but cause compile warnings for Ryan and Aaron, who for whatever reason
have a different default charset. Fixed.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3203 348d0f76-0448-11de-a6fe-93d51630548a
from a stream (see Vitter 1985). Thanks to Eric and Andrey for the pointer.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3197 348d0f76-0448-11de-a6fe-93d51630548a
sorting and merging interval lists, and creating RODs from intervals. This
gives Doug the ability to keep using our interval list parsing code when
sorting intervals on our behalf.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3159 348d0f76-0448-11de-a6fe-93d51630548a
Most of the changes occur in GenomeAnalysisEngine.java and GenomeLocParser.java:
-- parseIntervalRegion and parseGenomeLocs combined into parseIntervalArguments
-- initializeIntervals modified
-- some helper functions deprecated for cleanliness
Includes new set of unit tests, GenomeAnalysisEngineTest.java
New restrictions:
-- all interval arguments are now checked to be on the reference contig
-- all interval files must have one of the following extensions: .picard, .bed, .list, .intervals, .interval_list
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3106 348d0f76-0448-11de-a6fe-93d51630548a
a caveat: for anyone asking for all of the ROD's back from the RefMetaDataTracker (if your not using the facilities to get the track by name), you'll now be getting back a collection of GATKFeature objects. This object will contain the track name, and a method for getting the underlying object (getUnderlyingObject()), which will be the traditional RodVCF, rodDbSNP, etc. This layer is needed so we can integrate Tribble tracks (which don't natively have names). Calls that ask for RODs by name will still get back the traditional reference ordered data objects (RodVCF, rodDbSNP, etc).
Sorry for the inconvenience! More changes to come, but this is by far the largest (as has the greatest effect on end users).
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3104 348d0f76-0448-11de-a6fe-93d51630548a
of public interfaces. This won't be the last Picard patch, but it should be the last big one.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3096 348d0f76-0448-11de-a6fe-93d51630548a
Picard patch, including move of some of the Picard private classes we use to Picard public.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3087 348d0f76-0448-11de-a6fe-93d51630548a
quicker runtime. Making change as minimal as possible to avoid conflicts
with BT's incoming patch.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3061 348d0f76-0448-11de-a6fe-93d51630548a
This is *NOT* an ideal implementation. One day when we have lots of free time (or a greater desire), we will implement this correctly and sophisticatedly using all the power of JEXL. For now, though, this will have to do.
Docs coming tonight.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3060 348d0f76-0448-11de-a6fe-93d51630548a
1. Annotations can now be "decorated" with any arbitrary interface description - not just standard or experimental.
2. Users can now not only specify specific annotations to use, but also the interface names from #1. Any number of them can be specified, e.g. -G Standard -G Experimental -A RankSumTest.
3. These same arguments can be used with the Unified Genotyper for when it calls into the Annotator.
4. There are now two types of annotations: those that are applied to the INFO field and those that are applied to specific genotypes (the FORMAT field) in the VCF (however, I haven't implemented any of these latter annotations just yet; coming soon).
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3029 348d0f76-0448-11de-a6fe-93d51630548a
now do the minimum validation rather than the more rigorous only-within-the-contig-bounds
header validation.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3027 348d0f76-0448-11de-a6fe-93d51630548a
+ -L all now walks over all intervals
+ if a -L argument is passed with a .list extension, and file does not exist, returns a \
File Not Found error instead of "bad interval" error. We plan to soon revisit interval \
lists and generate a concrete list of filenames, so this is likely temporary.
+ Error is thrown if the start position on an interval is higher number than the end position.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3021 348d0f76-0448-11de-a6fe-93d51630548a
-Check that base and qual strings are the same lengths
-Fix one more bug in the clipper.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3006 348d0f76-0448-11de-a6fe-93d51630548a
(stop - start is length-1 on closed intervals, so we need to check greater than OR equals to zero)
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2990 348d0f76-0448-11de-a6fe-93d51630548a
1. emit AA,AB,BB likelihoods in the FORMAT field for Mark
2. remove constraint that genotype alleles (in the GT field) need to be lexigraphically sorted.
3. Add bam file(s) used by genotyper to header for Kiran
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2963 348d0f76-0448-11de-a6fe-93d51630548a
Removing obsolete genotyping classes.
First stage of removing dependence on old Genotype class.
More changes to come.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2960 348d0f76-0448-11de-a6fe-93d51630548a
We almost completely support indels. Not yet done with plink stuff.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2926 348d0f76-0448-11de-a6fe-93d51630548a
2. Significant refactoring of Plink code to work in the rods and use VariantContext. More coming.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2913 348d0f76-0448-11de-a6fe-93d51630548a
simplification of some of the locus traversal code.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2886 348d0f76-0448-11de-a6fe-93d51630548a
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()
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2809 348d0f76-0448-11de-a6fe-93d51630548a
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).
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2773 348d0f76-0448-11de-a6fe-93d51630548a
VCFRecord - "." dbsnp-ID entries now taken into account (thought these were represented as null; but I guess not)
VCFGenotypeRecord - added a replaceFormat option; since intersecting Broad/BC call sets required genotype formats also be intersected (no changing on-the-fly)
VCFCombine - altered doc to instruct user to give complete priority list (was throwing exception if not)
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2760 348d0f76-0448-11de-a6fe-93d51630548a
1) Now cares about Genotype filtering. If it is flagged as filtered, it can count as a FP/FN/TP; but goes into a "non-confident genotype" bin, rather than het/hom.
2) Can give it a Genotype Confidence flag (-GC) which will automatically filter genotypes in the way above for quality > Q for "-GC Q"
3) Can give it an -assumeRef flag. For sites only in the truth VCF (that don't even appear in the variant VCF), that locus will be treated as confident
ref calls for all individuals in the variant VCF; and the calculators updated accordingly.
*** Important: Default behavior is that sites unique to the truth VCF are considered no-call sites for the variant. This flag can help get aroudn that;
however the safest way to run this is to have a variant VCF with calls at each and every locus, if that is possible.
VCFGenotypeRecord -- added an isFiltered() call to automate looking up the FILTERED flag for VCF v3.3
SimpleVCFIntersectWalker - basic outline for a walker I'm working on tonight.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2747 348d0f76-0448-11de-a6fe-93d51630548a
when you want to protect your VCF header from being infected by the samples in a bound hapmap VCF. Changes are as follows:
VCFRecord - minor change to adapt isNovel() to the case where the dbsnp ID field is empty, but the info field has DB=1
HapmapVCFRod - introduced for the reason at the top
RODRecordIterator - was: catch ( Exception e ) { throw new StingException("long ass message") }
is now: catch ( Exception e ) { throw new StingException("long ass message",e) }
to permit full stack ejaculation.
RodVCF - Now with more brackets!
ReferenceOrderedData - registering HapmapVCF as a bindable string
VariantAnnotator - There's an extra space on a line. And some new brackets.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2733 348d0f76-0448-11de-a6fe-93d51630548a
support batched intervals in a single shard, but intervals are not yet compressed into a single
shard.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2730 348d0f76-0448-11de-a6fe-93d51630548a
1) Changed VCF/RodVCF to allow for inquiries to whether or not the site is novel; isNovel() looks at the ID field, and those members of the info field that indicate membership in dbsnp, hapmap2, or hapmap3; and if none can be found, returns true.
2) Changed VariantAnnotator to annotate hapmap2 and hapmap3, if you bind rods to it with those names. Works in the same way as DBSNP does -- if you give it a rod named "hapmap2" it'll annotate membership in it. -- Passes integration tests
3) Changed UnifiedGenotyper to do the same thing (since it uses Annotations as a subroutine) -- Passes integration tests
4) Changed MultiSampleConcordanceWalker to take a flag --ignoreKnownSites (or -novels) to examine concordance only on sites that are not marked as in dbSNP or in Hapmap in the variant VCF
5) Changed VCFConcordanceCalculator (the object MultiSampleConcordanceWalker runs on) to output Concordant_Het_Calls and Concordant_Hom_Calls separately, rather than combined as Concordant_Calls
6) AlleleBalanceHistogramWalker -- I don't know what i did to this thing. I've been jerry rigging System.outs to do stuff it was never really intended to do; so there's probably some dumb System.out.print("HI I AM AT LOCUS:"+loc) stuck somewhere. It compiles at any rate.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2724 348d0f76-0448-11de-a6fe-93d51630548a
Cleaned up SW code and started moving over everything to use byte[] instead of String or char[].
Added a wrapper class for SAMFileWriter that allows for adding reads out of order.
Not even close to done, but I need to commit now to sync up with Andrey.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2712 348d0f76-0448-11de-a6fe-93d51630548a
more like a resource bundle at this point, changed it over to use the Java ResourceBundle
support classes.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2606 348d0f76-0448-11de-a6fe-93d51630548a
update some straggler packages to the new package-info.java format introduced in 1.5.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2604 348d0f76-0448-11de-a6fe-93d51630548a
* Moved GATKArgumentCollection into gatk.arguments folder to clean up the main folder, also added some associated argument classes (most of the changes).
* Added code the argument parsing system for default enums, we needed this so we could preserve the current unsafe flag, and at the same time allow finer grained control of unsafe operations. You can now specify:
"-U" (for all unsafe operations), "-U ALLOW_UNINDEXED_BAM" (only allow unindexed BAMs), "-U NO_READ_ORDER_VERIFICATION", etc.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2586 348d0f76-0448-11de-a6fe-93d51630548a
SequenomToVCF now correctly has no-calls when probes fail.
Re-enabled SequenomToVCF integration test.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2572 348d0f76-0448-11de-a6fe-93d51630548a
I've updated HapMap2VCF to use the new interface; Chris agreed to take care of Sequenom2VCF.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2558 348d0f76-0448-11de-a6fe-93d51630548a
SequenomToVCF - Takes a sequenom ped file and converts it to a VCF file with the proper metrics for QC. It's currently a rough draft,
but is working as expected on a test ped file, which is included as an integration test.
Modified:
VCFGenotypeCall -- added a cloneCall() method that returns a clone of the call
Hapmap2VCF -- removed a VCFGenotypeCall object that gets instantiated and modified but never used
(caused me all kinds of confusion when I was basing SequenomToVCF off of it)
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2554 348d0f76-0448-11de-a6fe-93d51630548a
Also, moved some code that pulls samples out of rods from VCFUtils into SampleUtils.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2552 348d0f76-0448-11de-a6fe-93d51630548a
-MQ0 annotation is now standard again
-Added AC and AN annotations to VCF output
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2545 348d0f76-0448-11de-a6fe-93d51630548a
-Added useful trim() method for Strings for characters other than whitespace
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2538 348d0f76-0448-11de-a6fe-93d51630548a
2. Adding a preliminary version of the new Genotype/Allele interface (putting it into refdata/ as the VariantContext really only applies to rods) with updates to VariantContext. This is by no means complete - further updates coming tomorrow.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2533 348d0f76-0448-11de-a6fe-93d51630548a
QualityUtils - Stole the BaseUtils code for flipping reads around and applied it to quality scores
SecondBaseSkew - Nothing's really different, just a commented line
Additions (experimental annotations for future development of second-base annotation)
** I DO NOT INTEND FOR ANYONE TO USE THESE **
- ProportionOfNonrefBasesSupportingSNP
- ProportionOfSNPSecondBasesSupportingRef
- ProportionOfRefSecondBasesSupportingSNP
+ I hope these are self-explanatory
- QualityAdjustedSecondBaseLod
+ Adjust lod-score by 10*log10[P[second bases are as observed]]
Added walker:
QualityScoreByStrand - oneoff project that's being saved if i ever need it
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2527 348d0f76-0448-11de-a6fe-93d51630548a
Returns list of Pairs <String,Integer>, where each pair consists of a unique indel event observed at the site and the total number of observations of that event. String representation for insertions is verbose (e.g. +ACT), while deletions are represented as "5D" (since read backed pileup has no reference information, so we can not get actual sequence of deleted bases)
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2479 348d0f76-0448-11de-a6fe-93d51630548a
This change allows the pooled calculation model to work correctly with multiple threads. Boys, the Genotyper is now officially parallelized.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2462 348d0f76-0448-11de-a6fe-93d51630548a
Also, array size for caches should be max score + 1.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2444 348d0f76-0448-11de-a6fe-93d51630548a
1: all overlapping and abutting intervals merged (ALL),
2: just overlapping, not abutting intervals (OVERLAPPING_ONLY),
3: no merging (NONE). This option is not currently allowed, it will throw an exception. Once we're more certain that unmerged lists are going to work in all cases in the GATK, we'll enable that.
The command line option is --interval_merging or -im
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2437 348d0f76-0448-11de-a6fe-93d51630548a
bthomas
ebanks
depristo
hanna
aaron
delangel
chartl
asivache
weisburd
rpoplin
kcibul
alecw
kshakir
andrewk