The translation of mRNA into peptide is carried out by the ribosome, made up of two parts (larger subunit and smaller subunit). There are three binding sites (E, P and A) for tRNA on the larger subunit. The process of translation consists of three main stages: initiation, elongation and termination. The initiation of translation is the process of a ribosome binding to the RBS (ribosome binding site, also known as SD sequence) in the 5' UTR of an mRNA and scanning along until it intercepts the appropriate start codon. Then the elongation stage begins, during which the peptide gets elongated as the ribosome moves along the coding sequence of the mRNA. Finally, the translation is terminated when the ribosome intercepts the stop codon and the peptide is therefore released. The translation rate can be interpreted in RiPS (RiPS = Ribosomes Per Second; RiPSmc = RiPS per mRNA copy). Similar to PoPS, RiPS defines the level of translation as the number of ribosome molecules that pass a point on mRNA each second, on a per mRNA copy basis.

    During elongation, the rate limiting step is waiting for the correct tRNA to pair, which is determined by the corresponding tRNA availability (Varenne et al., 1984). Hypothetically, the tRNA availability in a species is largely represented by its codon bias (non-random usage of synonymous codons). Such bias is mostly suggested to be formed by evolutionary forces. Data of codon bias in over 35,000 organisms can be easily found in codon usage database ( Furthermore, slow or rare codons (cognate tRNA of relatively low concentrations) are not uniformly distributed along the coding sequence, but rather frequently occur in small clusters, especially near the 5' initiation site (Phoenix, D. A. & Korotkov, E., 1997). Particularly, the first 30-50 codons from the initiation site comprise a slow zone (the 'ramp') that has been observed both computationally and experimentally among many genes (Kurt Fredrick and Michael Ibba., 2010).

    As codon usage varies dramatically among different species, the RiPS of the same coding sequence expressed in different organisms might be accordingly different. The biobricks iGEM teams are using each year had actually been obtained from various organisms, thus displaying various codon biases. It might be entirely possible that one biobrick which had been successfully tested to work as expected would unfortunately fail in a new project, simply because of the two different host organisms used in each experiment.

    In order to make biobricks universally applicable and the translation process more predictable, we began our journey in iGEM competition this summer. We first built a mathematical model based on the biological background of the translational process to predict RiPS of any given coding sequences from any organism. The perspective on codon bias is especially incorporated into our model. Then our software Bach is made to further apply our work in real lab researches and realize our ideas of unifying codon bias of different protein coding biobricks from different generic backgrounds.


  1. Kurt Fredrick and Michael Ibba., 2010. How the Sequence of a Gene Can Tune Its Translation. Cell 141, 227-229.
  2. Phoenix, D. A. & Korotkov, E., 1997. FEMS Microbiol. Lett., 155, 63-66.
  3. Varenne, S., Buc, J., Lloubes, R., Lazdunski, C., 1984. Translation is a non-uniform process—effect of transfer-RNA availability on the rate of elongation of nascent polypeptide-chains. J. Mol. Biol. 180, 549–576.