untitled

168 7 Sequence Similarity Searching Tools

E-value P-value

10 0.9999546

5 0.9932621

1 0.6321206

0.1 0.0951626

0.05 0.0487706

0.001 0.0009995 (about 0.001)

0.0001 0.0001000

Table 7.1 Comparison betweenE-values and their correspondingP-values.

the homologous sequences, it is necessary to compare them with each other

and to organize them to show important features such as highly conserved

regions and other subtle similarities. This is known asmultiple sequence align-

ment(MSA). We have already mentioned MSAs in section 5.4 where they are

used in the PROSITE database. MSAs are also used in some BLAST variants.

Summary

• The raw scoreSfor a search can be normalized so that the results of dif-
  ferent searches can be compared.


• The expectation valueEcan be used to test whether the result of a search
  has statistical significance.

7.4.4 BLAST Variants

Since the launch of the original version of BLAST, many BLAST variants have

been developed. The major variants are presented in the following.

bl2seq http://www.ncbi.nlm.nih.gov/blast/bl2seq/bl2.html

In many cases, biologists only want to compare two sequences, rather than

embarking on a time-consuming journey of a full database search. The BLAST

2 SEQUENCES program is specifically designed for pairwise comparisons of

DNA or amino acid sequences (Tatusova and Madden 1999).

PSI-BLAST bioinfo.bgu.ac.il/blast/psiblast_cs.html

Queries using PSSMs differ from queries using substitution matrices in some

important ways. Unlike substitution matrices, there are no default or stan-

dard PSSMs. In fact, the PSSM is an important part of the query itself.