Almost 20 years ago, the Human Genome Project verified something that greatly complicated the way we understand biology: that proteins are not derived from simple expression of the genetic codes in nucleic acid sequences. The human genome contains open reading frames that code for roughly 20,000 canonical proteins. Around 70,000 additional proteins can be created by using alternative splicing of the DNA code. However, to our best understanding, a single human cell contains at least 1,000,000 different forms of proteins.
The elegance of the central dogma of biology is that information can be transferred from DNA to RNA to protein directly, yet it is now augmented by an understanding that reality is much more complicated. The life cycle of a particular protein is constructed by overlapping parameters such as gene regulation, post-transcriptional and post-translational modifications, which are regulated by interactions with transcription factors, enzymes, carrier proteins, and other nucleic acids. In addition, the function of the same protein requires multitude of interactions between other proteins, small molecules, and metabolic markers. Complexity in biology comes not only from the complexity of the networks of interactions involved, but also in the variations in the parts of those networks. This compounded complexity creates limits to what can be accomplished via reductionist approaches to biology, as well as limits to the usefulness of naive mappings of engineering and computer science concepts into biology.
This understanding led to the creation of the field of systems biology, the study of emergent patterns from the dynamic complexity of biology rather than focus on a particular type of molecule, such as genomics, proteomics, or metabolomics. An early editorial in Science put it well: “the pluralism of causes and effects in biological networks is better addressed by observing, through quantitative measures, multiple components simultaneously, and by rigorous data integration with mathematical models.” This powerful concept offers a way forward to understand how observable traits come about in biological systems and the creation of a new approach to measurement: multiomics.
The current approach to multiomics studies is rooted in a scale up of single omics biological techniques. In the recent years, collections of large multiomics datasets have enabled population studies that captures markers for disease diagnosis, lifestyle, and environmental conditions to better inform individualized and precise medical treatments. One major factor is the challenge to integrate information from independent omics studies into actionable insights. The analysis of muliomics data requires complex mining of the relations among different biological process streams using maching learning methods and multiperspective analysis.
While tremendous strides in increasing throughput and reducing cost have enabled multiomics studies through collaborative research, integration of simultaneous measurements from multiple biological components is not currently possible. Platforms of the various omics techniques including genomics, proteomics, transcriptomics, metabolomics, and cytomics are used in coordinated studies to produce snapshots of data with an intention of creating a holistic system level understanding. These studies can be quite difficult, normally requiring months or years to accomplish. There are outstanding difficulties in experiment planning, data integration, and cost control when performing multiomics studies. Even with the heroic efforts of these studies, the dynamic nature of biology is not addressed by current approaches, and computer models are required to simulate dynamic responses between individual time points.
At the same time that multiomics was being developed as an experimental foundation for systems biology, nanoelectronics was being developed as a foundation to directly link digital electronics to active biological systems. The combination of a unifying multiomics analytical biosensor with machine learning algorithms would be a very effective tool for complex biological data collection and decisive interpretation. Of the various designs and materials demonstrated over the 20 years of this effort, graphene transistor-based sensors are uniquely positioned to provide the capability to detect multiple analytes via an integrative device enabling greater accessibility to an academic research lab on a wider commercial scale.