We believe this method to be preferable to read mapping, because longer sequences are aligned and more robust alignments are obtained.
Our alignment only included data generated from the 454 for in-group taxa, but including several outgroup taxa with full-length complete data made for much more robust alignments and phylogenetic estimates.
This problem does not apply to the reads aligning to the contigs because their alignment positions to the reference genome can be inferred from the more robust contig alignments.
To achieve a more robust assignment of origin across the taxonomic breadth of this study, one would need to produce multiple sequence alignment and phylogenetic trees for each of the 90,440 transcripts in the E. lineata transcriptome.
While these issues are being addressed, genomic pursuits in zebrafish can focus on modalities that are more robust to nuances in alignment, such as genomic copy number changes and transcriptome profiles based on RNA-seq.
While their method can be considered more robust when good sequence alignments are available, we adopt the approach described here so that all yeast sequences may be included in our analysis.
Simulation results show that the incorporation of the restart scheme can make our CSRW-based alignment method more robust, especially when the available topological data are either unreliable or insufficient for detecting the similarities between networks (see Section S2).
Since different multiple alignment software packages can yield very different alignments, use of the profile HMM approach is more robust than relying on a single multiple alignment to construct a consensus or a position-specific score matrix.
We would like to add a word of caution regarding consensus phylogenies: there is a body of evidence indicating that the "supermatrix" approach (concatenated alignments) provides more robust phylogenetic reconstructions compared to the "supertree" (consensus) methods[ 72- 75].
Searches were initially run with a stringent e-value threshold of 1e-35 until convergence (to obtain confident alignments and more robust profiles) and later the threshold was relaxed to the value corresponding to 0.01 times the score of the first false positive and searches were continued until convergence.
