Design Powerful Predictor For MRNA Subcellular Location Prediction In Homo Sapiens

Messenger RNAs(mRNAs)shoulder special responsibilities that transmit genetic code from DNA to discrete locations in the cellular.There are countless mRNAs have been documented to accumulate in specific subcellular compartments.The concentrations of mRNA have been demonstrated to be related to cell polarity,cell motility,embryo development and other biological regulatory mechanisms.Thus,by identifying the mRNA location within cellular might provide spatial and temporal regulation of mRNA and protein functions.RNA fluorescent in situ hybridization and high-throughput subcellular RNA sequencing has been widely used to identify the presence and location of cellular nucleic acids.Although those traditional experiments have been successfully used in a variety of settings and provided lots of robust data,the experiments are both time-consuming and expensive.It is highly desired to develop computational tools for timely and effectively predicting mRNA subcellular location.In this thesis,a powerful predictor was developed to identify the mRNA subcellular localization in Homo sapiens(H.sapiens).The mRNA subcellular locations in H.sapiens with experimental evidence were retrieved from RNALocate and corresponding mRNA sequences were downloaded from GenBank.Assuming that the information gained from the sample with fewer labels is more focused,the samples that have already been detected one location were retrieved.The nonamer nucleotide composition of mRNA sequences was extracted to formulating mRNA samples.In order to eliminate the noisy and redundant information,two-steps feature selection with binomial distribution and one-way analysis of variance was performed.Finally,the optimal nonamer nucleotide composition contains 29745 features was obtained to constructed an mRNA subcellular location classifier based on support vector machine.In 5-fold cross-validation,results showed that the overall accuracy is 90.12% for H.sapiens.The predictor may provide a reference for the study of mRNA localization mechanisms and mRNA translocation strategies.An online web server was established based on our models,which is available at http://lin-group.cn/server/iLoc-mRNA/.