Stämm

Roman Ortega Bianchi

Understanding cell state: a single Omic falls short?

People thought Genomics would solve everything. All of our diseases, disorders, agricultural challenges, evolutionary quandaries, ecological issues, even death itself. It didn´t.

The finalization of the Human Genome Project (HGP) in 2003 after 13 years of research, meant the beginning of a revolution in digital biology. Human genes were now discovered and accessible. That was it, we could know every mutation, insertion, deletion or duplication.

The industry brought down the cost of sequencing a whole human genome faster than Moore’s law could predict. From hundreds of millions of dollars to approximately six hundred in 2023.

The insights obtained by Genomics were fundamental for medical research and healthcare, allowing the development of personalized medicine, cancer genomics, rare genetic disorders, and pharmacogenomics. On the broader picture, its usefulness was proven in agriculture with crop improvement, bolstered nowadays by methods like CRISPR-CAS9.

However, several factors that Genomics couldn't explain hampered our understanding of living things. From expression patterns to epigenetic marks, knowing the DNA sequence alone did not provide all the answers. Genomics by itself doesn’t tell the cell’s complete story. That’s why a multiplicity of Omics of interest appeared: Proteomics, Metabolomics, Metagenomics, Phenomics, and Transcriptomics.

This is a new opportunity to understand cells like never before. Analyzing multiple Omics at the same time can be the game changer we were looking for with ambitious endeavors like the Human Genome Project.

Knowing and modifying DNA rapidly became a fundamental step in biology workflows. Bioengineering progressively became more accessible and cheaper, breaking down the bottleneck it used to be.

All these advancements allow us to raise the bar of the game and ask ourselves increasingly complex questions.

The development of cell lines for different industries has sped up and reduced its costs, especially for producing genetically modified clones. So, where are the bottlenecks? Understanding biological information in depth, specifically the relationship between phenotypes and ambient conditions.

The bottlenecks changed from data generation to data interpretation for decision-making optimization. Today, the main hardship is interpreting multiple Omics simultaneously while considering the environmental interactions that shape the ultimate phenotype. Plus, the insight loss of not analyzing the variability from exposing different cells to multiple physicochemical conditions and understanding how the phenotype would react.

Think, for example, of Diabetes. In most type 1 cases, people inherit risk factors from both parents, but no single gene causes the characteristic sustained high blood sugar levels. Researchers are studying potential environmental triggers such as weather, viruses or even early diet, showing that a genomic analysis would be insufficient.

The same is true for more than one type of cancer. Having a genetic predisposition is not the same as a diagnostic. That’s why taking into account multiple sources of information about patients and their tendencies can inform us about varying susceptibility to environmental triggers.

That is why Multi-Omics technologies enter the picture. By combining bioinformatics, computer science, and machine learning methods, we can benefit from the array of them available. It is not just a matter of evaluating base pairs. We must consider protein production, chaotic interactions, and associated specific metabolic pathways.

Biology involves the entire living world and is, therefore, infinite. DNA sequencing and other biological techniques will continue to increase the number and complexity of gene data sets. Having a Multi-Omics database is more than a matter of scientific curiosity. It is a necessary repository in case you truly want to understand what goes on inside living entities.

That is why researchers need Artificial intelligence (AI) or machine learning-based computational tools that can handle, extract, and interpret the valuable information hidden within this trove of data. In this way, mathematical models for AI would bring us closer to a global understanding of how environmental conditions affect cell output.

Multi-Omic data can fuel AI models, surpassing the results of analyzing a single Omic. Each provides variability that a machine learning model could learn to simulate unseen states of a cell before going to the lab.

Processing Multi-Omic data with AI saves time, reveals new patterns in the data that are unobservable by the naked eye, allows searching for biomarkers, finding new cell subtypes, understanding the performance of different cell lines in bioprocesses, or studying cell differentiation in stem cells.

With Multi-Omics analysis platforms, we could abandon plain metaphors that conceive DNA as the blueprint of living things to more accurate models, far removed from educational oversimplifications.

A new field of opportunity is created due to Multi-Omics analysis platforms, increasing the speed and amount of data analysis. Many laboratories can benefit from this type of advance to generate more discoveries and speed up decision-making in trials.

These advances have a relevant impact on academia and industry, two sectors constantly feeding back into each other. Whatever sector you are in, one thing is clear today: a single Omic falls short of understanding the complexity of life.