Fundamentals of Bioinformatics - Ogurtsov, A.N. 2013

Foundations of Bioinformatics
Examples of Data Comparison
Searching for Similar Sequences in Databases

Before performing biological sequence analysis, it is necessary to retrieve these sequences from Databases. For instance, if you have identified a sequence for a novel Gene or found a disease-causing gene in The Human Genome, you might want to know whether similar genes exist in other species. An ideal method is one that is, on the one hand, sensitive (capable of detecting even distant relationships) and, on the other hand, selective (ensuring that all identified relationships are genuine).

Database search Methods involve a trade-off between sensitivity and selectivity. Does the method find all or most of the sequences that actually exist, or does it miss a large portion of them? Furthermore, how many of the results returned by the method are false positives?

Suppose a database contains 1,000 globin sequences. Assume that a globin search in this database yields 900 hits, 700 of which are indeed globin sequences, while 200 are not. Such a search can be said to have 300 false negatives (missed, undetected sequences) and 200 false positives (detected sequences that are not actually the ones being sought). By lowering the threshold, we obtain fewer false negatives but more false positives.

It is often preferable to work with lower thresholds to ensure that nothing potentially important is missed; however, this requires a detailed inspection of the results to eliminate false hits.

A powerful tool for searching sequence databases using a query sequence is BLAST (Basic Linear Alignment Sequence Tool), which can be accessed via the NCBI website at http://www.ncbi.nlm.nih.gov/ (Figure 30).

Class="center">

Figure 30 - NCBI webpage; the link to the BLAST program is shown in the lower-right corner of the figure

Following the "BLAST" link at http://blast.ncbi.nlm.nih.gov/Blast.cgi is illustrated in Figure 31. The figure displays the portion of the page dedicated exclusively to basic subprograms. It includes:

✵ nucleotide blast - searches a nucleotide query sequence against nucleotide sequence databases using the blastn, megablast, and dmegablast (discontiguous megablast) algorithms;

✵ protein blast - searches an amino acid query sequence against protein databases using the blastp, psi-blast, and phi-blast algorithms;

✵ blastx - translates a nucleotide query sequence in all six reading frames and compares it against a protein sequence database;

✵ tblastn - compares a protein query sequence against a translated nucleotide sequence database;

✵ tblastx - translates a nucleotide query sequence in all six reading frames and compares it against a translated nucleotide sequence database.

Figure 31 - BLAST program webpage

Where:

megablast - a fast alignment tool optimized for finding highly similar sequences;

dmegablast - a fast alignment tool designed to find more divergent sequences with low similarity;

blastn - a traditional alignment tool for finding all similar nucleotide sequences;

blastp - a traditional alignment tool for finding all similar protein sequences;

psi-blast - Position-Specific Iterated BLAST - designed to detect distant evolutionary relationships;

phi-blast - Pattern Hit Initiated BLAST - searches for Proteins containing a user-specified pattern.

Pattern - (from English pattern - a sample, template, or model) refers either to a sequence motif or (less commonly) to a standard set of Procedures applied to different objects.

Example 4. Homologs of the human PAX-6 gene.

PAX-6 genes control eye development across a wide range of species.

The human, fly, and octopus eyes differ significantly in Structure. Previously, given the distinct competitive advantage provided by Vision, it was believed that eyes evolved independently in each evolutionary Lineage. Therefore, it came as a major surprise that the gene controlling human eye development has a homolog that governs eye development in Drosophila.

The PAX6 gene was first cloned in mice and humans. It serves as the master regulatory gene controlling the complex cascade of events in eye development.

Mutations in the human gene cause a clinical condition known as aniridia—an eye development defect characterized by an absent or malformed iris.

The homolog of the PAX6 gene in Drosophila is called the eyeless gene, which shares a similar function in controlling eye development. Flies with mutations in this gene develop without eyes; conversely, the expression of this gene on a fly's leg or antenna induces ectopic (i.e., misplaced) eyes. The eyeless mutant Drosophila was first described in 1915, and at the time, no one suspected its evolutionary relationship with the mammalian gene. The insect and mammalian genes are similar not only in sequence; they are so closely related that their activity transcends species boundaries. The expression of mouse PAX6 in a fly triggers ectopic eye development, just like the fly's own eyeless gene.

PAX6 homologs are also present in other phyla, including Flatworms, ascidians, sea urchins, and nematodes. The observation that rhodopsins (a family of proteins containing retinal as a chromophore) function as photosensitive pigments across various classes of organisms provides additional Evidence for the common origin of diverse photoreceptor systems.

The actual structural differences in the macroscopic Organization of various eyes reflect divergence and the independent evolution of highly complex structures.

The human PAX6 gene encodes a protein with the Swiss-Prot identifier P26367. This identifier can be retrieved by typing "PAX6" into the search bar of the UniProt program (Figure 23) and clicking "Search".

From the "BLAST" program window at http://blast.ncbi.nlm.nih.gov/Blast.cgi (Figure 31), we launch "protein blast" and enter the sp|P26367 identifier into the "Enter Query Sequence" box (Figure 32).

We select the "PSI-BLAST" algorithm in the "Program Selection" panel and initiate the search by clicking the "BLAST" button in the bottom-left corner of the window (Figure 32).

The search result is presented as an extensive window, the majority of which is occupied by a list of database entries matching the query sequence, sorted in descending order of statistical significance.

The beginning of this list is shown in Figure 33.

Figure 32 — BLAST input window

Figure 33 — BLAST output window

Each row contains a single match with a specific gene. Let us examine, for example, the third row from the top, which presents the results for:

The first Column, "Accession", contains the gene identifier (NP_001035735.1). This is a homolog of the Paired box protein Pax-6 [Bos taurus]. The databases are indicated in the final column; in this case, they are UniGene, Gene Structure, and Map Viewer.

The number 868 represents the score assigned to the detected alignment match.

The statistical significance of this match (E-value) is measured as E = 0.0.

The E-value (expectation value) is defined as the probability that a given degree of similarity could arise by chance.

The E-value represents the expected number of sequences that would match the query as well as or better than the target sequence by chance when searching a database of the same size containing random sequences.

E = 0.0 indicates a complete identity.

Alignment parameters will be discussed in detail in Section 8.5.

The pairwise alignment result between the human PAX6 gene and the bovine PAX6 gene (located at the bottom of the same window below the match list) demonstrates their absolute similarity (Figure 34).

Figure 34 - Pairwise alignment of human and buffalo PAX-6 genes in the BLAST output window

The higher the E-value, the greater the divergence in sequence comparison. For instance, for the Drosophila twin eyeless homolog (NP_524638.3), the E-value is 7∙10^-145 (Figure 35(a)).

Figure 35 - Similar human and Drosophila twin eyeless PAX-6 genes: a - search result; b - pairwise alignment in the BLAST output window

The pairwise alignment of the human PAX-6 and Drosophila twin eyeless genes reveals significant sequence differences (Figure 35(b)).

METABOLISM/35.html">Review Questions and Exercises

1. List the hierarchical Levels of biological nomenclature using humans and fruit flies as an example.

2. What are homologous Organs?

3. What is the difference between Divergent and convergent evolution?

4. Into what three domains did Carl Woese divide all organisms based on ribosomal RNA analysis?

5. What symbols are used in the ClustalW2 results window to denote: identical Amino Acids, similar amino acids, gaps, and lack of sequence similarity?

6. What is the difference between sequence similarity and Homology?

7. What are false negative and false positive search results?

8. What is the purpose of the BLAST program?

9. What is nucleotide blast?

10. List the algorithms used in nucleotide blast.

11. What is protein blast?

12. List the algorithms used in protein blast.

13. What is blastx?

14. What is tblastn?

15. What is tblastx?

16. What is megablast?

17. What is psi-blast?

18. WHAT IS A pattern?

19. What is the E-value (expectation value) of sequences?