Harnessing Multi-Omic Data with Deep Learning
We are delighted to welcome DeepLife to the Tetra Partner Network and to share a recent conversation with Jonathan Baptista, Co-Founder & CEO. As integrative multi-omic data grows exponentially, scientists are seeking the ability to effectively generate actionable insights. DeepLife’s Discovery Platform utilizes Deep Learning to leverage multi-omic data so that customers can gain more scientific insights, faster.
In a sentence or two, please tell us about DeepLife.
DeepLife is a next-generation systems biology company harnessing omics data and the power of Deep Learning to model cells and efficiently engineer their behavior, accurately predicting cell reactions to various perturbations. DeepLife’s proprietary “Discovery Platform”, which includes its “Virtual Cell” product, uses novel approaches based on state-of-the-art multi-omics data, Deep Learning, and systems engineering to identify molecular triggers that drive a human cell back to its healthy state.
What customer needs does your Discovery Platform solve?
Human cells are the most advanced pieces of engineering on earth, encompassing trillions of chemical interactions associated with normal, healthy physiology. Small perturbations within these interactions, due to mutations or external influences, can lead cells to pathological states and ultimately to disease. The massive scale and complexity of the inner workings of human cells has traditionally impeded scientists’ efforts to identify the drivers of diseases through the iterative reconstruction of cell mechanisms. Furthermore, the exponential growth of integrative multi-omic data being produced outstrips the capacity of scientists to effectively generate insights from this data.
The exponential growth of integrative multi-omic data being produced outstrips the capacity of scientists to effectively generate insights from this data.
DeepLife solves this insight deficit problem in two ways through its proprietary DeepLife Discovery Platform, which leverages multi-omic data to shine a light on the inner workings of the human cell, thereby generating the actionable insights scientists need to deliver a sorely needed step change in the speed and quality of development of biological assets.
First, DeepLife’s Discovery Platform utilizes a robust Deep Learning - based engine for ingestion of massive amounts of single cell data, as well as for creation of curated, harmonized, and annotated single cell data atlases of tissues and diseases. The atlases serve as the basis for DeepLife’s single-cell ‘Analysis’ offering. DeepLife Analysis supports large-scale multi omics data analysis workflows to give scientists and engineers who analyze drug discovery data a scalable ecosystem designed around the intense demands of the omics era.
Second, the platform features its Deep Learning - based Virtual Cell capability for generating a systems biology level model of how a human cell functions in normal and diseased states; thus streamlining target/biomarker identification and validation, mechanism of action understanding, drug repurposing, and many kinds of in-silico modeling.
Systems Biology has long been a Holy Grail. How does your platform deliver on that promise?
Systems Biology has indeed been a Holy Grail for biologists over the years. Despite the cell’s complexity, powerful computational/mathematical analysis and modeling of complex biological systems has been developed and deployed, leading to new insights into underlying molecular physiology of the human cell. Also, the proliferation of multi-omic data has made it possible to infer missing parts of a comprehensive picture of an unhealthy cell.
DeepLife’s use of this multi-omic data to create digital twins of cells in silico that enables it to deliver on the promise of systems biology and to profoundly improve drug discovery productivity.
It is DeepLife’s use of this multi-omic data to create digital twins of cells in silico that enables it to deliver on the promise of systems biology and to profoundly improve drug discovery productivity. DeepLife’s Virtual Cell offering provides comprehensive pathway analysis and supports the identification and prioritization of druggable targets and biomarkers for disease specific models.
How do you handle the deluge of data coming from NGS?
We at DeepLife say “the more data, the better”. DeepLife’s Discovery Platform applies a Deep Learning - based data curation pipeline to data taken from customer’s sequencing instruments, as well as from public sources, to efficiently ingest and structure genomic, transcriptomic, and other data. The pipeline curates data from all sources, helping bioinformaticians to cope with the rapid growth of proprietary and public omics data, as well as with the variability of quality of the data.
The more data, the better.
What’s the value of the DeepLife Datastore?
The DeepLife curation pipeline outputs structured public omics cell, tissue, and organ atlases that are available in our “Datastore” product. This marketplace provides users with one-click access to Deep Learning - ready multi-omic data sets, as well as application programming interface (API) access to interoperable data packages. DeepLife has mapped more than a dozen atlases with more than 20 million single cells, spanning brain, blood, liver, lung, intestine, and other tissues and organs. These FAIR (Findable, Accessible, Interoperable, Reusable) - compliant atlases are updated on a monthly basis, drawing data from public data repositories and scientific articles.
DeepLife has mapped more than a dozen atlases with more than 20 million single cells, spanning brain, blood, liver, lung, intestine, and other tissues and organs.
How much data is needed to build a digital twin?
The more data that is provided, the better the digital twin model becomes. Because of this, we have built the DeepLife Discovery Platform to ingest and analyze all forms of multi-omic data, including genomic, epigenomic, transcriptomic, and proteomic data. Furthermore, the platform processes data to support multi-model analyses of all kinds, as well as perturbation analyses (ie. CRISPR analyses).
DeepLife’s cloud-based platform has no IT setup and provides a tailored visualization coding environment and graphical interface, empowering users to run large-scale analyses on millions of cells, and to seamlessly integrate state-of-the-art interactive analysis tools using R and Python.
What are the benefits of belonging to the Tetra Partner Network?
The DeepLife Discovery Platform is a multi-omic data analytics and modeling solution that can process limitless amounts of omics data coming from the The Tetra Scientific Data Cloud™. This complementarity enables enhanced insight generation for client scientists in the areas of mechanism of action determination, target and biomarker advancement, and drug repurposing.
Scientists can combine TetraScience’s capabilities in management and harmonization of all types of scientific data with DeepLife’s analysis and modeling capabilities to improve the speed and efficiency of their discovery research activities.
What else would you like to share with us?
Using its DeepLife Discovery Platform and related omic data management and analysis services, DeepLife has been improving research outcomes for discovery scientists in the life sciences industry.
DeepLife has been improving research outcomes for discovery scientists in the life sciences industry.
We welcome discussions with potential partners for co-use of DeepLife’s Virtual Cell technology for enhanced biologic insight generation. You can read more about our technology in our recent Nature Communications publication.
Read the TetraScience / DeepLife press release.