Scores for the matches are calculated from the simple frequency matrix. It is the sum of scores at each position of the matrix.
A 'simple frequency matrix' is simply a count of the number of times any particular amino acid occurs at each position in the alignment used to create it. Simple frequency matrices are created using the program prophecy with the option '-type F' to create the correct type of matrix. The alignment should not have gaps in it.
The resulting matrix is moved to each position in the sequence(s) you are searching. At each position in the sequence, the frequencies of the amino acids or bases covered by the length of the matrix is read from the matrix. The sum of these frequencies at each position of the matrix is the score for that position of the sequence. If this score is above the threshold percentage of the maximum possible score for that matrix, then a hit is reported.
This is the ungapped aligned set of sequences used to make the matrix:
% more m.seq >one DEVGGEALGRLLVVYPWTQR >two DEVGREALGRLLVVYPWTQR >three DEVGGEALGRILVVYPWTQR >four DEVGGEAAGRVLVVYPWTQR% prophecy Creates matrices/profiles from multiple alignments Input sequence set: m.seq Profile type F : Frequency G : Gribskov H : Henikoff Select type [F]: Enter a name for the profile [mymatrix]: Enter threshold reporting percentage [75]: Output file [outfile.prophecy]:
|
The ouput is a list of three columns.
The first column is the name of the matching sequence found.
The second is the start position in the sequence of the match.
The third column (after the word 'Percentage:') is the percentage of the
maximum possible score (sum of the highest value at each
position in the frequency matrix).