Long Non-coding RNA And Environmental Factors Associations Predictive Research Based On Gaussian Interaction Profile Kernel Similarity

The difference between the inherent nature of biological individuals and the behavioral habits developed after their birth is mainly caused by genetic and environmental differences.Phenotypes and diseases are determined by the complex interaction of genetic factors(GFs)and EFs(Environmental Factors,EFs).According to the central dogma of molecular biology,genetic information is mainly stored in the DNA sequence.Genetic information is transcribed from DNA to RNA,and then translated from RNA to protein.RNA is the intermediate between DNA sequence and its encoded protein.Genome sequence analysis showed that in the human genome,the sequence of the encoded protein accounts for less than 2% of the DNA sequence,and the remaining 98% of the DNA sequences do not encode proteins.There is a wide variety of non-coding RNAs in biology.Among them,non-coding RNA molecules with a length of 200-100000 nt are called Long non-coding RNA(LncRNA).LncRNA is an important noncoding RNA that regulates the life process.There are numerous studies on the association of protein-coding genes and microRNAs with EFs,while fewer methods based on bioinformatics and computation to study the association between LncRNA and EFs.The work of this paper is to study the prediction of the correlation between LncRNA and EFs based on the nuclear similarity of Gaussian interaction profile(GIP)kernel similarity.By using bioinformatics and available biological data,a reliable calculation method for predicting the relationship between LncRNA and EFs is studied to solve the blindness and high cost problem of biological experiment methods,which plays a guiding role in the experimental research of biologists.The main content is as follows:(1)Based on the known data of the association between LncRNA and EFs,a similarity matrix of LncRNAs and EFs was obtained by introducing the Gaussian interaction profile kernel similarity method for the later algorithm.(2)Using the Gaussian interaction profile kernel similarity of LncRNAs and EFs,three computational models for predicting the association of LncRNAs and EFs were proposed.They are the predictive methods of LncRNA and EFs association based on Laplacian regularized least square method,the prediction method of LncRNA and EFs association based on KATZ method,and the prediction method of LncRNA and EFs association by bi-random walk algorithm on the heterogeneous network.(3)This paper mainly designs and implements the above three algorithms,analyzes and compares the performance of each method to predict the relationship between LncRNA and EFs,and validates the reliability of the algorithm through experiments.Finally,a comprehensive comparison and analysis are made between the prediction results of the method designed in this paper and the existing methods.The results show that the prediction methods presented in this paper all have high credibility and reliability.