I was looking at this paper which talks about the miscalculations in the BLOSUM matrices and how the "wrong" BLOSUM62 matrix performs better than the corrected version. I was trying to find the corrected version they talk about to experiment with, but the link they list in the paper appears dead.

Does anyone know where I can find the corrected matrix that the authors talk about? (preferably a version listing frequencies rather than scores)

  • 2
    $\begingroup$ Either write to them asking for the files or see the images in the supplementary info (fig 4-5). They could have at least made it as a table; unfortunately you have to manually type it out. $\endgroup$
    Sep 2, 2014 at 4:30

2 Answers 2


Okay, I emailed Dr. Mark Styczynsk (one of the authors) to ask about a link to the matrix. Here are the links he gave me:



http://web.mit.edu/bamel/blosum/RBLOSUM64 (isentropic with BLOSUM62)

http://web.mit.edu/bamel/blosum/RBLOSUM62 (not isentropic, but it's 62)

The first link is to the c code to generate the corrected BLOSUM. According to Mark, just replacing the original c file in the original BLOSUM tarball with the revised c file can generate the corrected matrix for any of the BLOSUMs (30, 50, 92). Haven't yet tried that myself.


Your question made it difficult to understand what you are looking for so I hope these two links will provide you with the desired data.

0. A source link to the paper in question and its supplementary file.

1. What blast is currently using.

2. What the math calls for.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .