Open image in new window Figure 6 Comparison of the number of retained and discarded positions for two datasets.
For the grid dataset (Fig. 6A), 3B yielded 36,756 discarded positions and 36,311 retained positions, resulting in a total of 73,067 positions with a discard rate of 50.3%.
In contrast, FID3B yielded 28,036 discarded positions and 61,248 retained positions, resulting in 89,284 positions in total with a much lower discard rate of 31.4%.
We ran both 3B and FID3B with the two datasets mentioned above for the same 320 iterations and recorded the numbers of retained and discarded positions (Fig. 6).
For the gradient grid dataset (Fig. 6B), 3B yielded 40,792 discarded positions and 31,938 retained positions, resulting in a total of 72,730 positions with a discard rate of 56.1%.
We explain how our method affects the reconstruction results based on two aspects: the selection of the initial positions of fluorophores and the number of retained and discarded positions of fluorophores.
We also discarded positions where only a single sequence was found and, in the interspecific analyses, assembly regions where sequences from only one of the species assayed were detected.
Using only the first and second codon position (i.e., discarding positions that could had been subjected to multiple substitutions; see above), the topologies and resolutions reflected those obtained for all codon positions (data not shown).
Because these coupled "signals" may be masked by the "noises" generated by other positions, we checked each position and perturbation before clustering, and discarded 47 positions and 4 perturbations for which the standard deviation of their coupled values is less than 0.1 in corresponding rows or columns [34].
In the next step, we follow to position 10 (at index 1); a quick check for the gap range [0,0] discards position 13.
