*contributed equallySynonymous codon choice can have dramatic effects on ribosome speed, RNA stability, and protein expression. Ribosome profiling experiments have underscored that ribosomes do not move uniformly along mRNAs, exposing a need for models of coding sequences that capture the full range of empirically observed variation. We present a method, Iχnos, that models this variation in translation elongation using a feedforward neural network to predict the translation elongation rate at each codon as a function of its sequence neighborhood. Our approach revealed sequence features affecting translation elongation and quantified the impact of large technical biases in ribosome profiling. We applied our model to design synonymous variants of a fluorescent protein spanning the range of possible translation speeds predicted with our model. We found that levels of the fluorescent protein in yeast closely tracked the predicted translation speeds across their full range. We therefore demonstrate that our model captures information determining translation dynamics in vivo , and that control of translation elongation alone is sufficient to produce large, quantitative differences in protein output.As the ribosome moves along a transcript, it encounters diverse codons, tRNAs, and amino acids. This diversity affects translation elongation and, ultimately, gene expression. For instance, exogenous gene expression can be seriously hampered by a mismatch between the choice of synonymous codons and the availability of tRNAs. The consequences of endogenous variation in codon use have been more elusive, but new methods have revealed that changes in translation speed due to synonymous coding mutations, upregulation of tRNAs, or mutations within tRNAs can have dramatic effects on protein expression, folding, or stability [1][2][3] . However, translation initiation has been considered the rate-limiting step in translation, implying that changes in elongation speed should have limited effects 4 . Recent work has suggested a relationship between codon use and RNA stability; slower translation may destabilize mRNAs . CC-BY 4.0 International license peer-reviewed) is the author/funder. It is made available under a The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/201517 doi: bioRxiv preprint first posted online Oct. 11, 2017; and thus decrease protein expression 5,6 . These opposing viewpoints have yet to be fully reconciled, leaving us with an incomplete understanding of what defines a favorable sequence for translation.With the advent of high-throughput methods to measure translation elongation in vivo , we can understand the functional implications of codon usage. Ribosome profiling measures translation transcriptome-wide by capturing and sequencing the regions of mRNA protected within ribosomes, called ribosome footprints 7 . Each footprint reflects the position of an individual ribosome on a transcript, and we can reliably infer the A site codon -the site of tRNA decoding -in each footprint (F...