This column shows the blastp E-value (expectation value) between each gene
and the selected gene. The greater the similarity of two proteins, the lower
the E-value is. Identical long proteins have an E-value of zero.
Formally, an E-value is the number of other known genes that are
expected to have at least this level of homology by chance. An E-value of
less than 0.1 can be safely interpreted as the probability that a match
this good would occur merely by chance.
Method
All known splice variants of all known human genes are
put into a blast protein database using the formatdb program.
This database is then searched with each known gene using
version 2.2.5 of the NCBI blastall program with the following
flags:
-p blastp -e 0.01 -m 8 -b 1000
These flags tell the program to do a protein-based search and report
up to 1000 hits of E-value 0.01 or better in a tab-delimited
format. The output of this search is loaded into the Gene Sorter database.
Action
Click on a gene's E-value to display an alignment between the gene and the
selected (highlighted) gene.