BamUtil: validate

From Genome Analysis Wiki
Revision as of 18:27, 21 July 2010 by Mktrost (talk | contribs)
Jump to navigationJump to search

Status

The initial version of a SAM/BAM Validator is complete. It can be found at: http://www.sph.umich.edu/csg/mktrost/bam/

Purpose

The BamValidator processes the specified SAM/BAM file:

  1. to determine if it has any syntactic or format violations.
  2. to generate basic statistics.

The user can then decide if they want to use the file for future processing based on whether it passed syntactic/format validation and based on the statistics that were reported.


Valid SAM/BAM File Requirements

A valid SAM/BAM file meets the validation criteria specified in SAM Validation Criteria.

Statistic Generation

The statistics only reflect alignments that were successfully read from the BAM file. Alignments that failed to parse from the file are not reflected in the statistics, but alignments that are invalid for other reasons may show up in the statistics.

The following Statistics are generated by the BAM Validator if the --disableStatistics option is not set:

Read Counts
Statistic Description
TotalReads Total number of alignments that were successfully read from the file.
MappedReads Total number of alignments that were successfully read from the file with FLAG bit 0x004 set to 0 (not unmapped).
PairedReads Total number of alignments that were successfully read from the file with FLAG bit 0x001 set to 1 (paired).
ProperPair Total number of alignments that were successfully read from the file with FLAG bits 0x001 set to 1 (paired) AND 0x002 (proper pair).
DuplicateReads Total number of alignments that were successfully read from the file with FLAG bit 0x400 set to 1 (PCR or optical duplicate).
QCFailureReads Total number of alignments that were successfully read from the file with FLAG bit 0x200 set to 1 (failed quality checks).
Statistic Description
MappingRate(%) 100 * MappedReads/TotalReads
PairedReads(%) 100 * PairedReads/TotalReads
ProperPair(%) 100 * ProperPair/TotalReads
DupRate(%) 100 * DuplicateReads/TotalReads
QCFailRate(%) 100 * QCFailureReads/TotalReads
Statistic Description
TotalBases Sum of the SEQ lengths for all alignments that were successfully read from the file.
BasesInMappedReads Sum of the SEQ lengths for all alignments that were successfully read from the file with FLAG bit 0x004 set to 0 (not unmapped).

NOTE: If the TotalReads is greater than 10^6, then the Read Counts and Base Counts specify the total counts divided by 10^6. This is indicated in the output with a (e6) appended to the field name.


How to Use the Bam Validator Executable

Parameters

	Required Parameters:
		--in : the SAM/BAM file to be validated
	Optional Parameters:
		--noeof             : do not expect an EOF block on a bam file.
		--so_flag           : validate the file is sorted based on the header's @HD SO flag.
		--so_coord          : validate the file is sorted based on the coordinate.
		--so_query          : validate the file is sorted based on the query name.
		--maxErrors         : Number of records with errors/invalids to allow before quiting.
		                      -1 (default) indicates to not quit until the entire file is validated.
		                      0 indicates not to read/validate anything.
		--verbose           : Print specific error details rather than just a summary
		--printableErrors   : Maximum number of records with errors to print the details of
		                      before suppressing them when in verbose (defaults to 100)
		--disableStatistics : Turn off statistic generation

Usage

	./bam validate --in <inputFile> [--noeof] [--so_flag|--so_coord|--so_query] [--maxErrors <numErrors>] [--verbose] [--printableErrors <numReportedErrors>] [--disableStatistics]

Recommended Usage

If you don't want the file statistics, use --disableStatistics.

If you want to validate that the file is sorted, use the appropriate sorting flag. If you trust the @HD SO flag, use so_flag, otherwise if you want to check that it is sorted by coordinate, use --so_coord.

If you want to see the error details, use --verbose, but if you want to limit the number of errors displayed, use --printableErrors.

If you just want to know if the file is validly formatted or not, use --maxErrors 1

The following will give the most information (without validating that the file is sorted): 

./bam validate --in <inputFile> --verbose


Return Value

  • 0: all records are successfully read, are valid, and are properly sorted.
  • non-0: at least one record was not successfully read, not valid, or not properly sorted.

Example Outputs

Valid File

./bam validate --in ~/data/bamExample/37mer_alt.bwa.bam

The following parameters are available.  Ones with "[]" are in effect:

Input Parameters
 --in [/home/mktrost/data/bamExample/37mer_alt.bwa.bam], --noeof,
               --maxErrors [-1], --verbose, --printableErrors [100],
               --disableStatistics
   SortOrder : --so_flag, --so_coord, --so_query

'
Number of records read = 18900000
Number of valid records = 18900000

TotalReads(e6)	18.90
MappedReads(e6)	14.77
PairedReads(e6)	18.90
ProperPair(e6)	11.28
DuplicateReads(e6)	0.00
QCFailureReads(e6)	0.00

MappingRate(%)	78.17
PairedReads(%)	100.00
ProperPair(%)	59.68
DupRate(%)	0.00
QCFailRate(%)	0.00

TotalBases(e6)	699.30
BasesInMappedReads(e6)	546.67
Returning: 0 (SUCCESS)

Invalid File

./bam validate --in test/testFiles/testInvalid.sam 

The following parameters are available.  Ones with "[]" are in effect:

Input Parameters
 --in [test/testFiles/testInvalid.sam], --noeof, --maxErrors [-1], --verbose,
               --printableErrors [100], --disableStatistics
   SortOrder : --so_flag, --so_coord, --so_query


Number of records read = 32
Number of valid records = 2

Error Counts:
	FAIL_PARSE: 17
	INVALID: 1
	INVALID_QNAME: 3
	INVALID_RNAME: 8
	INVALID_POS: 2
	INVALID_CIGAR: 2
	INVALID_QUAL: 2

TotalReads	14.00
MappedReads	14.00
PairedReads	6.00
ProperPair	0.00
DuplicateReads	0.00
QCFailureReads	0.00

MappingRate(%)	100.00
PairedReads(%)	42.86
ProperPair(%)	0.00
DupRate(%)	0.00
QCFailRate(%)	0.00

TotalBases	47.00
BasesInMappedReads	47.00
Returning: 7 (INVALID)

Invalid File with Verbose

Printable errors is specified to produce a smaller example that does not print all the errors since that would take up more space. 

./bam validate --in test/testFiles/testInvalid.sam --verbose --printableErrors 5

The following parameters are available.  Ones with "[]" are in effect:

Input Parameters
 --in [test/testFiles/testInvalid.sam], --noeof, --maxErrors [-1],
               --verbose [ON], --printableErrors [5], --disableStatistics
   SortOrder : --so_flag, --so_coord, --so_query

Record 1
INVALID_QNAME (ERROR) : Invalid Query Name - the string length (256) does not match the specified query name length (0).
INVALID_QNAME (WARNING) : Invalid Query Name (QNAME) length: 256.  Length with the terminating null must be between 2 & 255.

Record 2
INVALID: 0 length Query Name.

Record 3
INVALID_QNAME (WARNING) : Invalid character in the Query Name (QNAME): ' ' at position 2.

Record 4
FAIL_PARSE: flag, 29M5I3M:F:295, is not an integer.
FAIL_PARSE: Invalid Tag Format: *, should be cc:c:x*.

Record 5
FAIL_PARSE: Too few columns (1) in the Record, expected at least 11.


Number of records read = 32
Number of valid records = 2

Error Counts:
	FAIL_PARSE: 17
	INVALID: 1
	INVALID_QNAME: 3
	INVALID_RNAME: 8
	INVALID_POS: 2
	INVALID_CIGAR: 2
	INVALID_QUAL: 2

TotalReads	14.00
MappedReads	14.00
PairedReads	6.00
ProperPair	0.00
DuplicateReads	0.00
QCFailureReads	0.00

MappingRate(%)	100.00
PairedReads(%)	42.86
ProperPair(%)	0.00
DupRate(%)	0.00
QCFailRate(%)	0.00

TotalBases	47.00
BasesInMappedReads	47.00
Returning: 7 (INVALID)


Libraries

Retrieved from "http://genome.sph.umich.edu/w/index.php?title=BamUtil:_validate&oldid=1801"