Research

Automatic Phylogenetic Footprint Analysis
using the available sequences of HOXa clusters as an example

• Don't think that you can get any useful information about parameter settings
  by reading the introduction if you have not used blastz before.
  
• Blastz works link a polymerase: ignores the outmost ends of the sequences
  If you blast two sequences, cut out a lokal alignment found and blast it again,
  you will find a lokal alignment at least 4(6)bp shorter than the original one.
  2(3)bp at the 5 prime end, 2(3)bp at the 3 prim end.
  
• good reasons to change the default B=2 If you want to change something but get the same output set B=1:
  it will compare the two sequences as they are AND
  sequence1 with the reverse complement of sequence2.
  If you are absolutly sure that your sequences are from the same strand
  you may set B=0 and save time by just comparing the sequences as they are.
  
• good reasons to change the default C=0
  C=0 means that any region of sequence1 can give a match with any region of sequenc2.
  If you think that having an alignment aligning position i to position j in sequence1
  with position k to position l in sequence2 an other alignment may just occure
  aligning j to 'end' with k to 'end' or 'begin' to i with 'begin' to k
  then set C=1 or C=2 (I am not yet sure that this makes any difference).
  Further I have to notice that the value of C has an influence on the score of an alignment.
  
• The misterious Threshold K
  If you want K just to reflect the sum of the partial scores
  be carefull with the value for Parameter C
  and set Parameter P to (0).
  
• Parameter P=1 takes into account the 'repetitivness' whithin the sequence
  Sequences with low entropy (high repetitivness) get a lower score then
  the same matches ordered in a random fashion (high entropy).
  Entropy is used by default (P=1) maybe you want to tern it off (P=0).