(3) is also outdated..
[Questions]
(i) In most alignments analyzed, SBP detected breakpoints based on AIC and AIC-c, while no recombination based on BIC. Would you use BIC to judge presence of recombination in alignments?
(ii) SBP may be affected by the number of sequences in alignments? (alignments must include at least four sequences to detect phylogenetic incongruence)
(iii) SBP may falsely detect recombination due to the presence of positive selection?
[References]
(i)
Multimedia File Viewing and Clickable Links are available for Registered Members only!! You need to
BIC was selected because it had the best power/accuracy performance in our initial simulation studies, comparing AIC [36], AIC-c [37] and BIC
(ii)
Multimedia File Viewing and Clickable Links are available for Registered Members only!! You need to
We noticed that when samples were reduced, the GARD algorithm found a significantly higher number of recombination breakpoints. For example, in the nef gene of HIV-2, when 24 sequences were used, only one breakpoint was inferred. When downsized to 10 sequences, four breakpoints were estimated.
(iii)
Multimedia File Viewing and Clickable Links are available for Registered Members only!! You need to
We note that large scale surveys for recombination using traditional detection software [e.g., [22]] are possible, but such programs likely suffer from an inverse effect, i.e. recombination may be falsely detected due to the presence of positive selection.