A likelihood ratio test was performed in order to differentiate highly significant domain‐GO term mappings from the ones that could be observed just by chance. 4.3.2 Computing E scores (likelihood ratio test) A new probability matrix, θ ̄ Di,GOj , was defined by running the EM algorithm again for each Di, GOj pair, under the model that θ ̄ Di , GOj Di , GOj is set to zero at the first step of every iteration.
← all excerpts
Mutual annotation-based prediction of protein domain functions with Domain2GO.
1
—
—