Translation initiation is a critical rate-limiting step in protein synthesis. It plays a pivotal role in determining cellular protein concentrations by setting an upper bound to the overall protein synthesis from a specific transcript. Factors such as mRNA structural stability, coding sequence length, and specific motifs influence initiation rates and collectively regulate in vivo translation of mRNA transcripts and thus the overall protein synthesis.
We developed a machine learning model that predict the in vivo translation initiation rates in S. cerevisiae transcripts using mRNA sequence features as input. The model was trained using in vivo initiation rates computed from wild-type S. cerevisiae transcripts. It accurately predicts the initiation rates for both endogenous and exogenous transcripts. The model is deployed as a web server – TIR predictor. It just required mRNA sequence length and, start and stop codon positions in an excel file. You can also just copy and paste these details manually to the web server.
We also developed a tool for gene optimization – TIR optimizer. This allows optimizing gene sequence (i.e., 5′ untranslated region and coding sequence) for a target initiation rate, that can vary between 0.04 and 0.19 per second. You can access the TIR predictor and optimizer via the following button.
The algorithm optimizes any given sequences for a specified initiation rate, considering nucleotide substitutions in UTRs and integrating codon and UTR substitutions in CDS and UTR regions. With practical applications in biotechnology and medicine, this work signifies a significant stride in advancing our understanding of translation initiation processes.
The link for the webserver is freely available in the link below.
Computational Biology and Biophysics Lab 