In parallel with the rapid growth of new sequencing technologies, many alignment programs [ 2- 20] have been developed, including MAQ, Novoalign (http://www.novocraft.com), SOAP, Bowtie, and BWA.
Over the last years, many alignment programs have been developed [ 67] to efficiently process millions of short reads and include, among others, Bowtie/Bowtie2 [ 68, 69], BWA [ 70, 71], MAQ [ 72], mrFAST [ 73], Novoalign (http://novocraft.com), SOAP [ 74], SSAHA2 [ 75], Stampy [ 76] and YOABS [ 77] (Supplementary Table S2).
In theory, sequence alignment cannot be considered separately from phylogenetic inference (e.g. many alignments programs use a guide tree), but both problems are NP-hard 26, 27 and in practice most researchers have regarded tree building as a distinct step (but see 28– 30).
The advances in sequencing technology over the last decade have brought new challenges in bioinformatics; consequently many new alignment programs that are much faster than BLAST have been published over the last few years.
To meet the requirement of efficient and accurate short read mapping, many new alignment programs have been developed.
When aligning disordered residues, only sequence information is used and all the predicted structure information is ignored; that is, their relevant feature values are set to 0. Unlike many sequence alignment programs (Altschul et al., 1997; Mott, 2005), CNFpred does not use an affine gap penalty.
The literature contains many short read alignment programs that have been benchmarked extensively [ 11, 12].
Dynamic programming algorithms, or a fast approximation, have been successfully applied to biological sequence comparison for decades, and this class of algorithms comprises the heart of many well-known sequence alignment programs (Batzoglou, 2005).
Many commonly-used multiple sequence alignment programs use the progressive sequence alignment strategy in which the alignment and the locations of insertions and deletions are permanently fixed in the growing alignment [21].
Previous attempts using multiple alignment programs to derive a phylogenetic tree across many eLRR subfamilies at once contain numerous differences from our results and from known relationships [ 3, 85].
According to structure superpositions, corresponding core regions in many homologous domains differ by fewer insertions and deletions than inferred by general alignment programs, reflecting the stability of the structural core of the protein family.
