Fig. 4: Model benchmark with proteins of biotechnological interest.
From: CodonTransformer: a multispecies codon optimizer using context-aware neural networks
Mean and standard deviation of Jaccard index (a), sequence similarity (b), and dynamic time warping (DWT) distance (c) between corresponding sequences for the 52 benchmark proteins across the 5 organisms (for organism-specific results, see Supplementary Figs. 20, 21, and 25, respectively). d Number of negative cis-elements in the 52 sequences generated by different tools for each organism (X shows the mean). Center line shows the median; box limits represent the 25th (Q1) and 75th (Q3) percentiles; whiskers extend to 1.5x IQR; points are outliers beyond whiskers. Data underlying this figure is provided in the Source Data File.
