hmm... Can I ask about the null models? I guess I thought the allignment would just pop out at the end... ++ Kevin Karplus-- > > One can write a simple HMM scoring program in a few lines, but it will be rather > slow. Training takes more lines. Doing sequence weighting take still more. > Doing space-efficient scoring with recovery of > alignment takes quite a few. Handling various different null models takes more > lines. Handling calibration of the E-values takes some more lines. > > After a while, it all adds up.