For AIC and BIC, both the best model and the 95% confidence set of models best fitting each alignment in the data set were considered in the analysis.
Alignments of syntenic noncoding loci were divided into SC and IR data sets based on their locations, then alignments in the two data sets were separately concatenated for further analyses.
These results suggest that the individual alignments in this data set are reasonably strong phylogenetic markers.
For a query domain, the p-value of a certain alignment with size s and core size c was defined as the fraction of alignments in the representative data set of size within 20% of s and core size exceeding c.
In these two simulated data sets, no matter whether the mismatch positions are designed to have high or low quality, all four aligners show a lower false alignment rate in the data set generated from 3000 exon regions (0.7-5%, see Table 9A, B) compared to the data set generated from 218 CpG islands that have more repetitive regions 14-177%, see Table 10A, B).
These SVRs correspond to 49% of the alignment positions in the data set.
ZZ carried out the data collection and sequence alignment, participated in the data analysis, and wrote the manuscript.
First, sequences designated Fragment, Partial, or Pseudogene were removed from the multiple sequence alignment resulting in the data set called "100" (all 3,286 complete tubulin sequences, including archaeal and bacterial "tubulin" genes, accounting for 2,633 alignment positions).
Table 1 shows the best-fitting model for the 287 RNA gene alignments in the EPO-35 data set.
We extract all human RNA sequences from the alignments in the Rfam "seed" data set (version 10.1), a total of 1,255 distinct sequences, associated with 550 Rfam families (Gardner et al. 2011).
