Research On Method Of Ontology Annotation And Knowledge Graph Construction Based On Multi-omics Data

With the development of sequencing technology,the sequencing cost has been reduced year by year,and lots of countries have developed large-scale precision medical plans.With the implementation of these large-scale precision medical plans,related biological data has been exploding.Currently,how to manage and analyze massive amounts of biological variants data is one of the huge problems facing Biomedical researchers.Although there are a lot of management software based on variants data,most of them are not combined with ontologies data.However,these ontologies information could not be ignored in genetic diseases and molecular diagnosis.The implementation of the precision medical plan is inseparable from the study of complex diseases.Complex diseases are a group of diseases caused by multiple genes or environmental factors.In the treatment of complex diseases,the analysis of single omics data is usually not enough,and a comprehensive understanding based on multi-omics knowledge is required.However,these multi-omics data are often stored in different databases,which is still an obstacle for biomedical researchers.Therefore,ontology annotation and knowledge graph construction based on multi-omics data is one of the important topics in the field of biomedicine.The main research results of this paper are as follows:(1)We have developed a sequencing analysis pipeline and an ontology annotation tool.This paper integrates the most popular mapping and variants calling software to complete the next generation sequencing(NGS),Third-generation sequencing and RNA-seq.And an ontology annotation tool has been developed based on the Variant Call Format(VCF).This tool could annotate ontologies information on variants files and integrate multiple databases into one file,which greatly improves query efficiency.(2)Knowledge graph construction method and semantic search model are developed based on multi-omics data.In this paper,we first construct the data pattern layer,and then build a knowledge graph based on this data pattern layer.The knowledge graph currently contains more than 300,000 nodes and more than 6 million relationships.Finally,in order to meet the researcher's semantic search needs,a semantic search model is constructed based on the knowledge graph.(3)An integrated platform has been Established for variants management and multi-omics based knowledge graph.It also includes a semantic search model based on knowledge graph.The platform uses B/S architecture,and the backend uses Mongo DB and Neo4 j databases.The front end uses a WEB interface to meet the researchers' variation management needs and semantic search needs,which is convenient for researchers.