Letter confidence (LC) is a measure of how consistently a letter x at a given position in a given sequence is aligned. It is calculated by comparing one MSA (the reference, R) to an ensemble of MSAs.
Let CR be the column of R where x is found. For each MSA A in the ensemble, let CA be the column of A where x is found, and CA be the column in A which has a majority of letters from CR. If CA = CB, then x is "well aligned", otherwise x falls in a different column from the majority. If no column in A has a majority of letters from CR, then x is not well aligned.
The LC of x is the fraction of MSAs in the ensemble where x is well aligned. The value ranges from zero to one.
See also column confidence.
Example showing how LC is calculated.
This shows a comparison between a reference MSA and one replicate in the ensemble.
The last columns match (this is not the only column in the example that matches).
In the last columns, YY.K is
found in both and these three positions in the reference is therefore well-aligned.
The G in the reference falls into a different column
and is therefore not well-aligned with this replicate.
