Modellentwicklung und maschinelles Lernen erhöhen die Proteinausbeute


Heterologous expression of genes requires their adaptation to the host
organism to achieve adequate protein synthesis rates. Typically
codons are adjusted to resemble those seen in highly expressed genes
of the host organism which lacks a deeper understanding of codon
optimality. The codon-specific elongation model (COSEM) identifies
optimal codon choices by simulating ribosome dynamics during mRNA
translation. COSEM is used in combination with machine learning techniques
to predict protein abundance and to optimize codon usage.