Research On Prediction Of Urine Excretory Proteins Based On Deep Learning

Disease biomarkers play important roles in detecting diseases as well as the mechanisms of disease occurrence and development.At present,the detection of biomarkers in body fluids such as blood,urine and saliva is an effective way to diagnose diseases.Because there are many signals for various physiological and pathophysiological conditions in the blood,most studies on body fluid biomarkers focus on blood.With the improvement of the proteomic analyses of urine samples in the last few years,researchers have found that urine is also an ideal source for detecting human diseases.Compared to blood,urine is a better source of biomarkers because the composition of urine is relatively simple and it can be easily and noninvasively obtained.Urine excretory proteins are among the most commonly used biomarkers in body fluids.Computational identification of urine excretory proteins can provide very useful information for identifying targeted disease biomarkers in urine by linking transcriptome or proteomics data.There are only a few methods based on conventional machine learning algorithms for predicting urine excretory proteins,and most of these methods strongly depend on the extraction of features from urine excretory proteins.The processes of feature engineering and feature selection may result in incomplete or biased features.In this paper,we proposed a novel deep-learning model for predicting urine excretory proteins.Different from existing urine excretory protein prediction methods,the model makes predictions directly based on amino acid sequences that are encoded as profile matrices by using the Position-Specific Iterative Basic Local AlignmentSearch Tool(PSI-BLAST).The framework of the model mainly consists of a convolutional neural network(CNN)module that can extract short motifs from input profiles,a recurrent neural network(RNN)module with Bi-long short-term memory(BLSTM)cells that can extract the long spatial dependencies between amino acids,and an attention module that assigns higher importance to amino acids that are relevant for the prediction.We validate that our model achieves a good accuracy on the training dataset and the independent test dataset(91.25% on the training dataset;88.98% on the independent test dataset).Our experiments also show that the proposed outperformed existing methods based on conventional machine learning models.The proposed model,combined with transcriptome and proteomics data,can provide very useful information for the detection of targeted disease biomarkers in urine.By comparing urinary protein biomarkers with our model results,we find that our model can achieve a true-positive rate of over 80% for urinary protein biomarkers that have been detected in more than one study.We also combine our model with transcriptome and proteomic data from lung cancer patients to predict the potential urinary protein biomarkers of lung cancer.A web server is developed for predicting urinary excretion proteins.We believe that our predictive model and web server are useful for biomedical researchers interested in identifying urinary protein biomarkers,especially when they have candidate proteins for analysing diseased tissues using transcriptome or proteomics data.