PSWN / PSSM - Position Specific Scoring Matrices

1) Start with data that was used to create consensus (here sequences thought to contain similar motifs)

2) Related sites are discovered using a search engine

3) Align sequences and then score each of the six positions

4) New sequences can gain score for who close they match = product of each score

PROBLEMS:1) Anything times by 0 = 02) Does not take into account G and C content

SOLUTION1) Pseudocounts2) Background frequency of particular nucleotide in genome

5) Final formula: Score (position, nucleotide) = (q+p)/(N +B)

Key:q = observed counts for the nucleotide at the given p = weighted pseudocounts = B * (overall frequency of nucleotide) = 0.1*(0.32) N = total number of sequences (= maximum number of observed counts) = 8 B = total number of allocated pseudocounts = 0.1

6) Total score = Product of score at each position

PSSM - What is it?

Question

Answer