Fig. 2

The workflow followed in this study from in silico design to lab bench scale production. Three design algorithms, OptKnock, OptGene and OptForce, were used on Yeast 8.5.0 to find gene candidates (Additional file 1: Table S5). 17 of these gene candidates were prioritised through further in silico predictions (Figs. 4, 5). After selecting the gene candidates, the yeast strains containing single genomic modifications were designed (Table 1). To engineer the yeast genome, the ACtivE Toolkit was used. The strains were then screened in deep-well plates to detect the good performance genomic modifications (Fig. 6). These modifications were then combined to design second-level yeast strains containing multiple genomic modifications (Table 2). The best-performing strains were then screened in the Biolector platform with real-time pH, DO and biomass monitoring (Fig. 7 and Fig. 8). Finally, the best-producer was used for the productions a lab bench scale bioreactor (Fig. 9)