What Will Happen If One Read Can Map To Two Regions?

Question

[microtan0902]

Supposing two regions are completely same in human genome. I know BWA will map the read to one of the two regions by random. but may I know what will do by bowtie, bowtie 2, and soap2? Are there any ref papers I can cite?

[Ashutosh Pandey]

In this case, all the aligners will give that read a very low mapping score. For example, Bowtie will give it a mapping score of 3. BWA will give it a mapping quality of 0. Additionally, aligners will also list out the alternative mapping position in the BAM file. For example, BWA will use "X0:Number of best hits" tag and put it as "X0:2" and use "XA Alternative hits; format: (chr,pos,CIGAR,NM;)" to give the location of other position.

Answer 1

[Mikael Huss]

In Bowtie and Bowtie2, you can control the behavior in this regard in a very detailed way, see e g the -a, -k, -m and other options. (http://bowtie-bio.sourceforge.net/manual.shtml and http://bowtie-bio.sourceforge.net/bowtie2/manual.shtml)

[Philipp Bayer]

Similarly in SOAP: -r 0 means don't report the read, -r 1 means report one of the locations (chosen at random), -r 2 means report all possible locations

Answer 2

[Varun Gupta]

For Bowtie and Bowtie2 you can control the behaviour of the output by not reporting all the alignments. The option -a would report all the possible alignments.

-a/--all
Report all valid alignments per read or pair (default: off). Validity of alignments is determined by the alignment policy (combined effects of -n, -v, -l, and -e). If more than one valid alignment exists and the --best and --strata options are specified, then only those alignments belonging to the best alignment "stratum" will be reported. Bowtie is designed to be very fast for small -k but bowtie can become significantly slower if -a/--all is specified. If you would like to use Bowtie with -a, consider building an index with a denser suffix-array sample, i.e. specify a smaller -o/--offrate when invoking bowtie-build for the relevant index (see the Performance tuning section for details)

By default bowtie reports 1 valid alignment per read.

-k <int>
Report up to <int> valid alignments per read or pair (default: 1). Validity of alignments is determined by the alignment policy (combined effects of -n, -v, -l, and -e). If more than one valid alignment exists and the --best and --strata options are specified, then only those alignments belonging to the best alignment "stratum" will be reported. Bowtie is designed to be very fast for small -k but bowtie can become significantly slower as -k increases. If you would like to use Bowtie for larger values of -k, consider building an index with a denser suffix-array sample, i.e. specify a smaller -o/--offrate when invoking bowtie-build for the relevant index (see the Performance tuning section for details).

Hope this helps

Varun