Changing Bowtie's Preferences

Bowtie

Bowtie has many preferences for aligning reads against reference sequences. The following preferences can be modified. If you are familiar with running Bowtie from the command line the command line parameter is shown in brackets along with the default value.

Note that the settings used to produce a Bowtie alignment are stored within the Comment Annotations field of the Reference Contig.

Bowtie2 uses less memory and is generally faster and more sensitive than Bowtie1. There is some advantage to using bowtie1 for reads less than 50bp. But nowadays the reads from all current sequencing technologies are much longer than this.

Here are the options that are exposed to the user in Assembler.

Miscellaneous

Bowtie2 will introduce gaps in both the reference and reads
All ambiguous bases in the reference are treated as Ns.
Bowtie2 does not use the strata concept at all which was used by Bowtie1
Bowtie2 uses an index of a reference sequence rather than a sequence itself. Assembler will generate the index automatically. Note that future versions of MacVector will allow you to save indices to increase the speed of alignments for subsequent alignments.
Hit reporting is no longer exposed to the user as its use is determined to some extent by the presets that are used. Now the default is to always report the best alignment.

Preset

Bowtie2 comes with eight sets of preset parameters. With speed versus accuracy. Note that the presets do change whether you are doing a local or global alignment. Again Assembler deals with this for you and just presents you with four presets:

very fast

fast

sensitive (default)

very sensitive

The full parameters are below. See the Bowtie manual for explanations of these.

--very-fast-local Same as: -D 5 -R 1 -N 0 -L 25 -i S,1,2.00

--fast-local Same as: -D 10 -R 2 -N 0 -L 22 -i S,1,1.75

--sensitive-local Same as: -D 15 -R 2 -N 0 -L 20 -i S,1,0.75

--very-sensitive-local Same as: -D 20 -R 3 -N 0 -L 20 -i S,1,0.50

--very-fast Same as: -D 5 -R 1 -N 0 -L 22 -i S,0,2.50

--fast Same as: -D 10 -R 2 -N 0 -L 22 -i S,0,2.50

--sensitive Same as: -D 15 -R 2 -L 22 -i S,1,1.15 (default in ?end-to-end mode)

--very-sensitive Same as: -D 20 -R 3 -N 0 -L 20 -i S,1,0.50

Type of Alignment (-local, default -local).

Bowtie2 has local alignment as well as global alignments (end to end). In local alignment mode it will trim bases from either end of a read if a better alignment is the result. Global (end to end) alignments will map the entire read against the reference, although do remember that gaps may be inserted (see below). The default is Local. This was changed in MacVector 15.5.

No gaps within first bases of read

Bowtie2 will now insert gaps in the reference and reads where doing so will increase the alignment score. Such INDELS are reported in the VCF tab and are also graphically represented in the MAP tab. By default gaps will not be introduced into the first four bases of a read.

Threads

The -p parameter defines how many threads are used. The default is 1. I suggest we increase this to 2.

It does increase speed, but obviously also memory requirement. Note in testing with -p 8 and an E.coli reference with 2.5 million Solexa paired reads I never saw total memory go above 0.5Gb.

Paired End Alignments

USE PAIRED-END ALIGNMENTS: Default is ON for projects containing two reads files and OFF for all others

MINIMUM INSERT SIZE (-i, Default: 0).The minimum insert size for valid paired-end alignments. For example if 60 is specified and a paired-end alignment consists of two 20-bp alignments in the appropriate orientation with a 20-bp gap between them, that alignment is considered valid. A 19-bp gap would not be valid in that case.

MAXIMUM INSERT SIZE (-X, Default: 250): The maximum insert size for valid paired-end alignments. For example if 100 is specified and a paired-end alignment consists of two 20-bp alignments in the proper orientation with a 60-bp gap between them, that alignment is considered valid. A 61-bp gap would not be valid in that case.

ORIENTATION:

Forward Reverse (fr): the paired reads are 5' to 3' at one end with 3' to 5' at the other end. i.e. the pair are on opposite strands and pointing to each other.
Forward Forward (ff): Both reads are 5' to 3' and on the same strand
Reverse Forward (rf): the paired reads are 3' to 5' at one end with 5' to 3' at the other end. i.e. the pair are on opposite strands and pointing away from each other.

Note: Most, but not all, Illumina paired end reads are -fr. The Mate Pair protocol for Illumina produces -rf and Solid reads are always -ff.

There are two new concepts with regards to pairs. Bowtie2 now has a mixed mode. If it cannot find a paired match, within the given parameters, for a supposedly paired read it will treat each pair mate as single reads. There is also the concept of concordant and discordant aligned pairs. Concordant means it matches all the given parameters. Discordant means that they still align but with much relaxed limits with regards to the pair's relationship. This is distinct from the mixed mode stated above. Discordant matches is the default used by MacVector and cannot be changed. Please contact Support if you would like to use this parameter in a future release.

If the two reads of a pair align and match all pair specific criteria it’s regarded as a concordant match. Criteria includes the insert size, orientation, etc..
If a concordant match cannot be found then Bowtie will relax the criteria somewhat and again try to align the two mates of the pair. If it then aligns then it’s a discordant match. Note the relaxed criteria does not include how each individual read aligns to the reference, but rather the two reads relationship to each other.
If the pair still does not align then Bowtie2 will try and match each mate of the pair as a single read. This is the mixed mode.