BLOSUM: diferenças entre revisões

Conteúdo apagado Conteúdo adicionado
Linha 10:
:<math>S_{ij}= \left( \frac{1}{\lambda} \right)\log{\left( \frac{p_{ij}}{q_i * q_j} \right)}</math>
 
Aqui, <math>p_{ij}</math> é a probabilidade de dois aminoácidos <math>i</math> e <math>j</math> estarem substituindo uns aos outros em uma seqüência homóloga, e <math>q_i</math> e <math>q_j</math> são as probabilidades de fundo de encontrar os aminoácidos <math>i</math> e <math>j</math> em qualquer seqüência de proteína de forma aleatória.
<!--
 
<!--
Here, <math>p_{ij}</math> is the probability of two amino acids <math>i</math> and <math>j</math> replacing each other in a homologous sequence, and <math>q_i</math> and <math>q_j</math> are the background probabilities of finding the amino acids <math>i</math> and <math>j</math> in any protein sequence at random. The factor <math>\lambda</math> is a scaling factor, set such that the matrix contains easily computable integer values.
... The factor <math>\lambda</math> is a scaling factor, set such that the matrix contains easily computable integer values.
 
An article in Nature Biotechnology<ref>{{cite journal| year=2008| journal=Nat. Biotech. | volume=26 | pages=274–275| title = BLOSUM62 miscalculations improve search performance | doi = 10.1038/nbt0308-274 | url=http://www.nature.com/nbt/journal/v26/n3/full/nbt0308-274.html | author = Mark P Styczynski | coauthors = Kyle L Jensen, Isidore Rigoutsos, Gregory Stephanopoulos| pmid=18327232| issue=3 }}</ref> revealed that the BLOSUM62 used for so many years as a standard is not exactly accurate according to the algorithm described by Henikoff and Henikoff.<ref name=henikoff /> Surprisingly, the miscalculated BLOSUM62 improves search performance.{{cn|date=September 2011}}