Computation method for annotating semantic similarity by gene

Abstract

The invention provides a computation method of gene annotation semantic similarity. The method establishes correlation between the gene and the gene body node through the gene body correlation file provided by the gene body association; then the semantic similarity of the gene body node is firstly computed; the gene annotation semantic similarity is computed finally according to the semantic similarity of the gene body node. The computation method has the advantages of computing the gene annotation semantic similarity automatically and in large quantity.

Description

technical field [0001] The invention relates to a method for calculating the semantic similarity of gene annotation, which belongs to the technical field of bioinformatics. Background technique [0002] Gene ontology (gene ontology, GO) is an important gene annotation database, and biologists often use online tools such as AmiGO and QuickGO to retrieve gene GO annotations. After obtaining the gene annotations, it is necessary to compare the similarity of the semantics of the gene annotations, that is, to examine whether the functions of some genes are similar, or whether some genes are jointly involved in the metabolic process of certain substances, etc. At present, comparing the similarity of genes is mainly done manually. Since biologists usually need to compare dozens or even hundreds of genes, it will be very time-consuming and labor-intensive to compare the similarity of so many genes manually. , and the manual comparison will also be affected by subjective factors. U...

AI Technical Summary

Problems solved by technology

Although Jiang et al. improved the Resnik's method by considering the influence of node depth on semantic similarity, but as long as it is based on Resnik's method, there are two shortcomings: first, this method is based on node counting, ignoring the " Belong to (is-a)" and "Part of (part-of)" two different effects on node similarity; second, the node semantic similarity calculated by Resnik's method cannot be reused, because two The semantic similarity of nodes is affected by other nodes in the set they are in

Wang's method has two disadvantages: First, it is difficult to determine the descending ratio according to the relationship layer. Wang suggested to take 0.8 and 0.6, which is actually very random, and the determination of descending comparison has a direct impact on the similarity of GO nodes; Second, the weight value of the ancestor point is represented by the maximum weight value, ignoring the influence of different paths on the similarity of GO nodes

[0006] Therefore, there is currently no method for accurately calculating the semantic similarity of gene annotations.

Images