Welcome to UTailoR
UTailoR can be used to predict the effect of mRNA 5'UTR sequences on protein translation efficiency(measured by mean ribosome loading), and then generate 5'UTR sequences with higher ribosome load based on the original sequence.
We constructed a deep learning model with Convolutional Neural Network(CNN) and Gate Recurrent Unit(GRU) to predict the effect of a given UTR sequence on ribosome loading. Then each UTR sequence is processed by an autoencoder that changes a small number of neucleotides in the original sequence to generate a sequence that is similar to the original sequence but has higher MRL score. Different from genetic algorithm-based optimization methods, our method minimally changes the original sequence, which is safer and faster.
Contact us
Liu Yu Department of Biomedical Informatics Peking University Health Science Center 38 Xueyuan Rd, Beijing 100191 China Email: 1810305222@pku.edu.cn