Among the various algorithms for aligning multiple sequences, we selected the alignment approach implemented in the PRANK program [16].
To select the alignment positions on the basis of both, high conservation across DHA1 and high CREi, the normalized CREi was further scaled with respect to the RENULL as given by Equation 2. This was referred to as CRES and represented as per the following equation.
When two alignments had the same coverage, we selected the alignment with the higher score.
If several alignments are possible, we select the alignment that minimizes the mismatch score.
As Cons selects the alignment with the highest amino acid sequence conservation, we expect lower d N and d N/d S values with this method than with the other methods.
Select the alignment to decide where to place the watermark on image.
We also selected the alignments between ESTs and genomic clusters.
As might be clear beforehand, scoring functions which do not directly take sequence structure information into account such as ROSETTA do perform badly on the task of selecting the best alignment-induced templates.
Custom scripts were used to assess all alignments and select the final alignment for each cluster using only the particular paralog that yielded the longest alignment after trimming, that is, caused the smallest number of ambiguously aligned positions in the alignment.
