These new MSA construction methods are generally evaluated using one or more alignment benchmarks, for example, BAliBASE [15], OxBench [29] or PREFAB [24], and it is clear that this benchmarking has had a positive effect on their development [30].
Bowtie reports, for each read, one or more alignment containing no more than a few mismatches (two, by default) in the 5′-most s bases of the read.
However, it should be noted that all methods have adjustable alignment parameters, and in most cases could be optimized for more sensitive alignment.
The image after segmentation has much cleaner signals for more precise alignment and further analysis.
The lower levels of sequence divergence within the subgroups allowed for more reliable alignment reconstruction.
This is similar for more traditional alignment-free methods that calculated distances based on k-mer frequencies.
Our results on simulated data let us expect an even higher performance for more sensitive alignments and appropriate multi-mapper handling.
The read-mapping algorithm reports one or more alignments for each read within a scoring threshold, commonly expressed as the minimal acceptable significance of the alignment, or the maximum acceptable number of differences between the read and the reference genome.
In addition, the following uniqueness criterion was imposed to make sure that the two best alignments were unambiguous: for a query with three or more alignments, the alignments ranking third or worse were compared with the second best alignment (constituting the "short" partner of the reconstructed transcript).
Alternatively, for more divergent sequences, alignment criteria may be relaxed in short pre-defined space where the background noise is significantly reduced compared to a genome scale search.
