Local sequence alignment, similarity

A sequence alignment is a way of determining the similarity between two strings. This is a classical question in computer science, and has an exact solution referred to as the Smith/Waterman alignment. Unfortunately, this exact solution can be slow when analyzing large sequences, and therefore, approximate methods, such as Basic Local Alignment Search Tool (BLAST), have been developed to identify very similar sequences. 

Proteomics is concerned with the analysis of the complete protein complements of genomes. Thus proteomics includes not only the identification and quantification of proteins, but also the determination of their localization, modifications, interactions, activities, and functions. This chapter focuses on protein sequences as the sources of biochemical information. Protein sequence databases are surveyed. Similarity search and sequence alignments using the Internet resources are described. 

There are two approaches of sequence alignments A global alignment compares similarity across the full stretch of sequences, while a local alignment searches for regions of similarity in parts of the sequences. 

Figure 11,4. ExPASy Proteomic tools. ExPASy server provides various tools for proteomic analysis which can be accessed from ExPASy Proteomic tools. These tools (locals or hyperlinks) include Protein identification and characterization, Translation from DNA sequences to protein sequences. Similarity searches, Pattern and profile searches, Post-translational modification prediction, Primary structure analysis, Secondary structure prediction, Tertiary structure inference, Transmembrane region detection, and Sequence alignment.

Gracy, J. Argos, P. (1999). Automated protein sequence database classification. I. Integration of compositional similarity search, local similarity search, and multiple sequence alignment. Bioinformatics 14,164-73. 

Database search programs like FASTA [18] or BLAST [19] have been optimised to detect evolutionary relationships between proteins, and are readily adequate for template recognition and (multiple) sequence alignment in cases where the sequence identity is over 25-30% [20], The general procedure is to assume next that the backbone of the model is identical to the one of the template structure and add the side chains onto it [21], although some difficulties may arise with insertions, deletions and local low similarity. 

Therefore, and also to compare the results with existing information, one may still want to search the database with the results obtained by de novo sequencing. The approach to follow is a BLAST search (Basic Local Alignment Search Tool, [133-134]). Sequence alignment is a powerful way to compare novel sequences with previously characterized sequences and search for similarities. The BLAST search is much slower than the database searches outlined in Ch. 17.6.2. A nice example is the de novo identification of the bacterium Shewanella putrefaciens proteome by 2D-GE, in-gel digestion, direct infusion nano-ESl, and nano-LC-MS [135]. The results were BLAST searched against the incomplete genome of the bacterium. 

For DNA sequences, (local) sequence similarity can be calculated at the nucleotide level, at the peptide level and at both levels with the mixed-alignment option as previously explained. If the peptide-level or mixed-alignment option is used, it is possible to translate sequence segments only at the forward strand or to have the program look at both, the forward strand and the reverse complement. 

Two different output options are available to indicate the degree of (relative) local sequence similarity in the alignment. 

Morgenstern, B., Freeh, K., Dress, A. W. M., and Werner, T. (1998) DIALIGN finding local similarities by multiple sequence alignment. Bioinformatics 14, 290-294. 

---