Similarity Analysis of HIV

The visual query filters can aid enormously in constructing interesting queries for a large alignment report. In our previous paper [6], we analyzed the Human Immunodeficiency Virus (HIV) sequence, and found an interesting region containing alignments between HIV and the Simian Immunodeficiency Virus (SIV). Here we demonstrate how a similar analysis can be performed with greater precision using visual query filters.

In addition to isolating that region, suppose we are interested in alignments that are longer than 100 residues, because we want to filter out short motifs and look for a long alignment based on weak similarity. So we know that the alignment should not be an exact replica, but rather a loose identity. Thus, we take the number of matching positions (percent identities) to be between 30 and 70 percent, and the number of acceptable substitutions (percent positives) in the alignment to be between 40 and 75 percent. In addition, we last examined this sequence in 1994, and want to look at new alignment information only. Lastly, we are looking for a relatively strong hit, so the score should at least 130, and the P-value should be better than 10^-22.

The visual query filter constructed for the above statement is shown in Figure 4 in the design section. The position, score, and P-value are mapped to the X, Y, and Z axes, respectively. The result appears in Figure 7. Instead of a scatter plot of points, we see that each point is expanded into a glyph, and most short cluttering alignments have been filtered out, making any subsequent navigation around the information space faster and focusing our attention on the data of interest.

  
Figure 7: A visual query filter result showing alignments between HIV and SIV.

The glyphs in this figure are useful because they show large concentrations of positive colors that may indicate conserved regions. Conserved regions are important because they suggest where structural motifs or functional domains could be present. Remember that green is used as a positive color, representing a good substitution between two residue pairs. The alignments have green regions flanking large yellow regions. These alignments are to SIV sequences, thus marking a region of difference between the HIV and the SIV sequence. Since SIV is similar to HIV but cannot infect humans, biologists are very interested in such differences.