Non-linearity of Metabolic Pathways Critically Influences the Choice of Machine Learning Model

Figure 1

Classification of metabolic pathway modeling methods according to their complexity and the year of first application in this field. The ellipse size is proportional to the occurrence of the method for pathway modeling in the literature. Three main groups are defined: knowledge-based model (Michaelis and Menten, ; Chance, ; Shapiro and Shapley, ; Garfinkel et al., ; Savageau, , ; Fell and Small, ; Hatzimanikatis and Bailey, ; Curto et al., ; Heijnen, ; Liebermeister et al., 2010), data-based model (Wu et al., ; Cuperlovic-Culf, ; Ajjolli Nagaraja et al., ; Zampieri et al., ; Zhang et al., ; Kim et al., 2020) and hybrid model (Wiechert et al., ; Drysch et al., ; Antoniewicz et al., ; Nöh et al., ; Leighty and Antoniewicz, ; Antoniewicz, ; Pan et al., ; Yousoff et al., ; Heckmann, ; Oyetunde et al., ; Zampieri et al., ; Lo-Thong et al., ; Rana et al., 2020). Linear methods are represented in gray and non-linear ones are in blue. Methods in bold and white are those evaluated in this study.

Source link

Back to top button