Furthermore the BLAT search algorithm of the UCSC genome browser was used to find the best alignments in the human genome http://genome.ucsc.edu/cgi-bin/hgBlat.
PRANK+F alignments resulted in lower average d N/d S ratio estimates than MAFFT alignments in the human branch (table 4) although the results were less clear for other branches (supplementary file S2: tables S4 and S5, Supplementary Material online).
Phylogenetic comparison of the alignment (−326/+202 in the human sequence) led, as expected, to rejection of the molecular clock hypothesis.
The resulting alignment covered −326/+202 in the human sequence.
This site, located at position 354 in the alignment (position 349 in the human sequence), corresponds either to lysine or glutamate and was not detected by TreeSAAP.
The primate-specific regions were then masked by excluding repeats, CpG islands, CG dinucleotide, and any other regions not included in the EPO alignment block or in the human genome.
However, to ensure that the corresponding human and mouse genomic regions were extracted, we extended this region if, in an alignment of the human and mouse gene sequences, the TSSs could not be align.
In this case, the variant glutamine is more conserved in the protein alignment than the human native arginine (11% Q vs. 8% R), making predictions difficult for methods that over-rely on alignments.
We investigate the relationship of dN and dS from two different perspectives: a longitudinal approach following specific orthologous genes across a broad evolutionary time frame and a cross-sectional approach exploring the relationship of dN and dS for all genes in every pairwise alignment with the human sequence.
The percentage of multiple mapping in zebrafish was 5-fold higher than in our previous Bismark alignment of the human RRBS library, where only 7.7% of the reads (75 bp length) showed multiple mapping against the whole human genome (GRCh37 build).
