Metabolite Discovery in Perspective

My first incredibly pleasant read this year: a perspective article “Metabolite discovery: Biochemistry’s scientific driver” by Martin Giera, Oscar Yanes and Gary Siuzdak (published in Cell Metabolism, January 4, 2022).

The publication is talking about metabolomics evolution and how it helps to characterize metabolites. Starting with the deep history journey to the 18th century and the characterization of urea, the article covers the progress of the metabolite identification across the centuries and discusses the development of novel technologies and tools.

The article sections are set as an unfolding historical novel, with the following titles.

-?????????1700s to 1900: the beginning and metabolite characterization in bulk

This part is a journey to the past which highlights its tight connection with the present. It depicts a development of our knowledge about cholesterol: from first discovery in 1758, to obtaining its name 50 years later, establishing its molecular formula in 1888, digging into its structure half a century later, and finally synthesizing it almost 200 years after it was first discovered.

-?????????1900s: analytical technology drives discovery

These years laid the foundation of the systems biology. The developments allowed for the convergence of the primary omics: genomics, proteomics, and metabolomics. Authors say that this convergence became possible with the bioinformatics emergence, as the ultimate “omic-glue”.

-?????????21st century knowns: the annotated road to known metabolite identification

Authors talk about the challenges in identifications from vast amounts of highly complex data, which can be compared to “the cocktail party problem” (the task of hearing a sound of interest masked by overlaying background noise).

-?????????21st century unknowns: the road less travelled

Another challenge is the characterization of unknown unknowns, which typically starts with the observation of a metabolic feature(s) dysregulated during certain experimental conditions (e.g. healthy vs diseased, wild type vs mutant, etc.) for which no putative identity is available.

-?????????21st century: the road ahead

This section highlights the importance of bioinformatic tools in metabolomics progress. Databases growth at a rapid pace, and in silico approaches taking their place in the puzzle: artificial intelligence (AI), machine learning (ML), deep learning (DL), natural language processing (NLP) – will they solve all of the existing problems?

-?????????MS, NMR, and orthogonal technologies

The advancement in separation and detection technologies are still required to solve the metabolic “black matter” goal.

“Although we cannot yet precisely establish how many unknowns remain to be characterized, it is evident that new technologies and metabolomic approaches have already discovered new metabolites, metabolic reactions, and unexpected metabolic fluxes that have important physiological relevance”.

In conclusion, authors call metabolites “master manipulators of biology” and look forward to further developments in the youngest of the primary omics – the metabolomics.