PWM search: serious bug found and fixed

[COTRASIF]

Andreas from Germany brought to our attention that the maximal
theoretical similarity score of 1.00 is never achieved when using PWM
search, even though perfect matches are present in the results file.

Short investigation has revealed a serious bug in COTRASIF PWM search
tool: the last position (column) of the submitted matrix is not
accounted for during the search.
This bug is now fixed.

All the previous results of the PWM search tool had wrong (decreased)
scores, and had higher tolerance for the varying ending nucleotide.
Actually, all the previous search results should be viewed as if
search had been performed with a one-column-shorter matrix.

Please accept our apologies for any inconveniences caused by this bug.

And thanks Andreas for noticing and then reporting the problem.

--
COTRASIF developers