4. Incorporation of constraint-based modeling and machine
learning
In recent years, the advances in high-throughput devices and the rapid
growth of omics data provide a unique opportunity to depict biological
samples at multiple layers. Omics data generated from high-throughput
technologies are big data that can be analyzed individually or inferred
as multi-omic relationships through ML algorithms to gain more
biological insight [109-111]. Common omics datasets include
genomics, transcriptomics, proteomics, and metabolomics produced from
DNA sequencing, RNA sequencing and microarrays, and mass spectrometry,
respectively [112]. Furthermore, fluxomics is an additional layer of
omics generated from CBM approaches and includes metabolic flux
distribution values, thus representing the metabolic phenotype
[113]. Consequently, the integration of ML (for omics and
multi-omics analysis) and CBM (for generating fluxomics) looks promising
for analyzing a biological system such as cellular metabolism. However,
the capabilities of this hybrid approach are just now being explored,
and more research in this area is needed. In this section, first, we
briefly summarize the ways that ML and CBM can be coupled. Next, the
most prominent studies aiming at analysis and optimization of
fermentation parameters by ML-CBM approaches are reviewed.