Open main menu

Genome Analysis Wiki β

LibStatGen: VCF

Revision as of 18:44, 28 March 2012 by Mktrost (talk | contribs) (Created page with 'Category:C++ Category:libStatGen Category:libStatGen VCF = Variant Call Format (VCF) = The documentation on the Variant Call Format (VCF) can be found at: http://ww…')
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)


Contents

Variant Call Format (VCF)

The documentation on the Variant Call Format (VCF) can be found at: http://www.1000genomes.org/wiki/Analysis/Variant%20Call%20Format/vcf-variant-call-format-version-41

Our APIs are currently in the initial test phase, and are not yet available for public use.

API for Reading VCF Files

You will use both an VcfFileReader and an VCfRecord for reading VCF files.

VcfFileReader

An instance of the VcfFileReader class is used to read VCF files.

include file

VcfFileReader is declared in VcfFileReader.h, so be sure to include that file.

#include "AspFileReader.h"

Opening the VCF File

open opens the specified file and by default throws an exception if it was not successfully opened.

There are two methods for opening a file for reading. One is standard while the other allows the specification of a subset of samples to keep.

Both methods take the filename to be read as well as a reference to the VcfHeader to be populated.

The standard method is:

    // Open the vcf file for reading.
    VcfFileReader reader;
    VcfFileHeader header;
    reader.open("vcfFileName.vcf", header);

Subsetting Samples

To select only a subset of samples to keep, when opening the file also specify the name of the file containing the names of the samples to keep and the delimiter separating the sample names (default is a new line, '\n').

    // Open the vcf file for reading.
    VcfFileReader reader;
    VcfFileHeader header;
    // Subset 1 is delimited by new lines, '\n'.
    reader.open("vcfFileName.vcf", header, "subsetFile1.txt");
    // Open the vcf file for reading.
    VcfFileReader reader;
    VcfFileHeader header;
    // Subset 2 is delimited by ';'
    reader.open("vcfFileName.vcf", header, "subsetFile2.txt", ';');

Reading a VCF Record

Once the file is open, get the next record using readRecord.

It returns true if a record was successfully found and false on EOF or an error. Typically on a reading error an exception is thrown.

readRecord takes a reference to a VcfRecord object as a parameter. When true is returned, the VcfRecord is updated with the next record.

If you only process one record at a time (are done with a record before reading the next one), you can loop until readRecord return false, reusing the same record for each call.

        VcfRecord record;
        while(reader.readRecord(record))
        {
            // Your record specific processing here.
        }
        // Done reading the file.

If any subsetting was specified when the file was opened, the record that is returned will have only the specified subset of samples in it.

When reading a record, you can also specify discard rules. Internal processing of readRecord will apply the discard rules on the record that was read and will continue reading records and not return from the method until a record is found that should not be discarded (or until the end of the file). True is returned if a record that should be kept was found, false if the end of the file was hit without finding a record to keep.

Specifying Discard Rules

When specifying the discard rules, you should use the constants found at the top of VcfFileReader.h

For Example (see the file for the current set of Discard Rules and the associated values):

    static const int DISCARD_NON_PHASED = 0x1;
    static const int DISCARD_MISSING_GT = 0x2;
    static const int DISCARD_FILTERED = 0x4;
    static const int DISCARD_MULTIPLE_ALTS = 0x8;

You can set the Discard using setDiscardRules(uint32_t). You can add additional rules to what has already been set by using addDiscardRules(uint32_t).

        reader.setDiscardRules(DISCARD_FILTERED);
        reader.addDiscardRules(DISCARD_NON_PHASED);

is the same as

        reader.setDiscardRules(DISCARD_FILTERED | DISCARD_NON_PHASED);

and discards reads that do not have PASS in the FILTER field and reads that have a genotype that is not phased or have no GT in the FORMAT fields.

Check if EOF was reached

To see if the End Of the File has been reached, use isEOF().

        reader.isEOF();

True is returned if the End of the File has been reached, false if not.

See How Many Files Were Read

To see the total number of records in the file that were read, including any that were discarded, use getNumRecords.

To see only the records in the file that were kept (not discarded) use getNumKeptRecords.

        std::cout << reader.getNumRecords() << " were found in the file\n."
                  << "but only " << reader.getNumKeptRecords() << " were processed\n.";


Closing the VCF File

close closes the opened VCF file.

    reader.close();

VcfRecord

VCF records are read from VCF files using readRecord.

Once you have the record, use methods from the VcfRecord to extract the desired information.