In one study [ 5], researchers clustered ESTs and then aligned these clusters to the genome, looking for alignments that crossed gene boundaries.
Some of incorrectly oriented sequences can be detected by looking for alignments in the original direction having multiple non-consensus splices, that become all-consensus, high-identity alignments when reversed and complemented.
In each optimization criterion, we choose a particular gain function and look for alignment A∗ maximizing the expected gain A ∗ = arg max A E A T [ G (A, A T ) | X, Y ].
First, we looked for alignment of the initial Methionine (Met) amino acid to a downstream Met in multiple proteins, which may point to the less common initial Met being an incorrect translation start position.
We looked for alignments between the larger potentially multimodular proteins in E. coli and smaller orthologous proteins that are homologous to only one of the modules.
With the weak version of our anchored-alignment approach, we are looking for an alignment Ali wich maximum score such that the union Ali ∪ Anc is consistent.
For local alignment of primary sequences (Altschul et al., 1990; Pearson and Lipman, 1988; Smith and Waterman, 1981), one is merely looking for an alignment of contiguous linear subsequences of a query and a target.
We start by estimating the average Rayleigh wave phase speeds over the whole network at each period by looking for phase alignments among band-pass-filtered CCFs in the time-distance domain (slant stack technique).
Therefore, A v gives the alignment value of each feature, and looking for the best alignment value for each feature, these values are sorted from the highest to the lowest.
While looking for an optimal alignment, the algorithm looks for certain constrained alignments based on what they call a hierarchical constraint.
