1. Substitution scores
In the Friday, Sep 9, lecture, we discussed pairwise alignments, and assumed
a score function s(a,b) for two bases a and b. The choice of this score 
function is a topic of research, particularly when we are aligning amino
acid sequences. Read DEKM Chapter 2.2 (page 14-15): the section on Substitution
matrices.

2. Affine gap scores in alignment
Another issue when scoring alignments is how to penalize gaps. The algorithm
we saw in class penalizes each gap independently, with a constant score. Often
alignment algorithms penalize a continuous sequence of gaps in an alignment
by the length of that gap sequence. There is a "gap initiation" penalty paid
when the sequence of gaps (in the alignment) starts, and thereafter each 
additional gap is given a constant "gap extension" penalty. Read how 
the dynamic programming algorithm is extended to handle such "affine gap 
penalties" in DEKM Chapter 2.4 (page 29): the section on "Alignment with 
affine gap scores".