Skip to main content
Figure 5 | BMC Bioinformatics

Figure 5

From: A jumping profile Hidden Markov Model and applications to recombination sites in HIV and HCV genomes

Figure 5

Reduction of active states for a set of HIV test sequences using the beam-search heuristics [32, 33]. In this example, we have 14 (sub-)subtypes each of which has three states per alignment column (match, delete and insert). Thus, a column corresponds to 14 × 3 = 42 states. The beam-search algorithm reduces the number of active states considerably; the figure indicates for each position in the query sequence those columns that contain active states. Thus, instead of considering the entire dynamic-programming matrix, our algorithm needs to consider only the small strip between the two lines. We used a beam width of MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBamrtHrhAL1wy0L2yHvtyaeHbnfgDOvwBHrxAJfwnaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaWaaeGaeaaakeaaimaacqWFSeIqaaa@377E@ = 10-20 and a jump probability of 10-9.

Back to article page