mRNA Lightning™ Platform Applications
Target Discovery
At the center of the platform is our mRNA knowledge graph that integrates knowledge from multiple proprietary sources: curated publications, our imaging dataset of over 2 billion mRNA visualizations, and the experimental results from executing thousands of actual experiments in our projects for over more than a decade of deep diving into mRNA biology. Unlike AI approaches that rely only on computational methods, mRNA Lightning actually visualizes disease biology with high-content high-throughput screens in disease model cells. It identifies dysregulated mRNA pathways and then proceeds to integrate this biological data with computational AI in a single framework: the LLM with the mRNA knowledge graph deep dive into proposed pathways whose modulation could potentially modify disease phenotypes. We then proceed to train our neural network to recognize the disease signature with millions of examples of candidate pathways, and the network decides which pathway illustrates the most distinct differences between healthy and diseased cells. This “disease signature” provides unprecedented insight into the underlying disease biology. Finally, the discovery of the mRNA neural image network is brought back into the knowledge graph to suggest targets related to the pathway. Targets are validated in actual disease models, using our dozens of mRNA MOA assays and imaging technologies. This unique process uses AI across the mRNA biology knowledge graph and the mRNA image neural network, combining computational and wet lab biology, leading to novel target discovery with direct biological evidence and a relevant link to the disease phenotype.
Unique advantages of our approach in Target Discovery
- AI across actual and computational biology: Enhances disease understanding through elucidation of disease signatures.
- Visualize disease biology: Provides new insights into underlying disease biology, with direct experimental evidence and a relevant link to the disease phenotype.
- Target discovery in mRNA biology: Identifies new targets with a clear understanding of their roles and mechanisms in mRNA biology.
Drug Discovery
Conceptually different from traditional phenotypic screening, our drug discovery leverages our ability to elucidate the disease signature, a dysregulated mRNA biology pathway that decodes the disease’s mechanism. Our high-content screening proceeds to identify disease signature modulators. Millions of biology experiments are conducted in our massively parallel automated mRNA biology lab with diseased cells perturbated by hundreds of thousands of small molecules. The millions of images of the signature pathway are uploaded to the cloud for analysis by the mRNA AI image neural network. The network identifies the active compounds, hits that have modified the pathway images taken from diseased cells so that they look like images of healthy ones. These disease signature modulators typically have multiple chemistries that share an inherent rationale for their MOA. Targeting the dysregulated pathway, their activity is highly selective to the disease mechanism. Our MOAi technology uses the LLM, the mRNA biology knowledge graph, and the Lightning co-pilot, enabling biologists to work hand-in-hand with the AI to sequence and execute dozens of mRNA biology MOA assays. Experimental results are brought back to the system and the process iterates to rapidly converge towards MOA and molecular target identification. Powered by the understanding of disease biology, compounds undergo rapid optimization against signature pathway biology and on-target chemistry.
Unique advantages of our approach in Drug Discovery
- Decoding disease biology and targeting its underlying mechanism: Traditional phenotypic screening approaches operate with no prior understanding of underlying disease biology or targets. As a result, it is very challenging to understand the mechanism of action of hit compounds and to identify their molecular targets. In contrast, our approach starts with diseased cells, elucidates a disease’s signature pathway, and screens for compounds that modulate underlying disease mechanisms. Hit compounds emerge out of the screen with an inherent MOA rationale, enabling rapid elucidation of molecular targets and optimization in the context of underlying disease biology.
- Multiple chemistries: Pathway modulators may have multiple chemistries, as they are potentially working through multiple targets along the pathway. This provides discovery projects with multiple shots on goal and fallback plans, lowering risk by increasing the probability of success.
- Compound selectivity: Multiple mechanisms regulate mRNA biology. By targeting the disease signature pathway, we are ensuring that molecules will have high efficacy and low side effects. The selectivity of such drugs is due to the fact that they bring diseased cells back to their healthy state through modulation of the specific pathway that tells them apart.
- Rapid optimization: Our process integrates data from computational and wet lab biology. In target discovery, we utilize AI in computational biology models, using the mRNA biology knowledge graph and our deep dive gene exploration analyses. Once hit compounds become available, our MOAi technology takes over to iterate between computational and wet lab biology in the sequencing and execution of dozens of MOA studies, using the Lightning co-pilot, LLM, and the mRNA image network to analyze results and rapidly converge to MOAs and targets.
Visual mRNA Vaccine and RNA-Based Drug Optimization
Analyzing the mRNA biology visualizations produced by our mRNA imaging technologies, the mRNA image neural network identifies a signature, the visual representation of the mRNA vaccine’s or RNA-based drug’s efficacy, observed within the cell’s biology. The network processes inputs of thousands of RNA constructs in a variety of cell types, visualizing and sorting them according to their efficacy signatures. It then compares these signatures, enabling novel strategies in vaccine or RNA-based drug design and optimization.
Visual mRNA Vaccines Optimization
mRNA Lightning is a game-changer in mRNA vaccine design and optimization, employing AI at scale for visual analysis of vaccine biology. We conduct rapid optimization cycles using a high-content, high-throughput platform, testing thousands of design variants across numerous cell types and conditions. Just like we have done in small molecule drug discovery, we are aiming to generate the world’s first visual dataset of mRNA vaccine biology, comprising over 1 billion images. This data will enable for the first time to understand vaccine biology at the cellular level, driving new strategies in vaccine design and optimization. The mRNA image neural network analyses images of mRNA biology and learns to identify visual vaccine efficacy signatures, which are crucial for ranking variants and guiding each iteration towards improved vaccine design, determining why certain designs succeed or fail. This vast amount of data feeds back into our ever-expanding visual dataset, improving neural network training and enriching our knowledge base for future projects. The approach incorporates the mRNA biology knowledge graph and a co-pilot with a natural language interface, enabling new understandings and strategies in vaccine design and development.
Unique advantages of our approach in mRNA vaccines optimization
- Visual vaccine optimization: Our technology visualizes mRNA vaccine biology inside cells at an unprecedented level of detail and scale. It then analyzes the millions of generated images, comparing different design variants across cellular conditions.
- Elucidating the mechanism of action of mRNA vaccines: A key advantage of our platform is its ability to elucidate the mechanism of action of mRNA vaccines through visual signatures. The neural network’s capability to extract a specific visual mRNA vaccine signature offers insights into the behavior of synthetic mRNA, such as its specific cellular localization or efficiency of delivery systems (e.g., LNP, ADC). This understanding of the mRNA’s journey and interaction within the cell is key for optimizing vaccine efficacy and delivery methods.
- Massively parallel, rapid optimization cycles: Our tera-scale mRNA biology lab operates a massively parallel infrastructure that runs high-content screens at high-throughput, allowing for rapid testing and iteration of hundreds of thousands of design variants.
Visual RNA-Based Drugs Optimization
Optimizing the efficacy, stability, and delivery of RNA-based drugs remains a challenge. mRNA Lightning offers a paradigm shift in making better and safer RNA-based drugs, employing AI at scale for visual analysis of the cellular biology of siRNA/ASO. Our massively parallel lab screens hundreds of thousands of siRNA/ASO designs across cell types and conditions, generating millions of images. These mRNA biology visualizations depict the biology of RNA-based drugs in cells, including translation, localization, decay, protein expression, siRNA/ASO-relevant machinery, and RNA-mediated stress responses among others. Our mRNA image neural network analyzes the images to visually recognize and extract the siRNA/ASO’s “biology signature”. Proprietary AI image analysis is then applied to compare the efficacy of designs, suggesting a subset for further development, along with the rationale for their optimization. This process is driven by our mRNA biology knowledge graph, our mRNA biology LLM, and the Lightning co-pilot.
Unique advantages of our approach in RNA-based drugs optimization
- Visualizing RNA drug biology: Offer unprecedented depth and scale in visual data, offering unique insights into siRNA/ASO behavior at the cellular level.
- Elucidating mechanisms of action of RNA-based drugs: The ability to extract specific "biology signatures" from visual data aids in comprehensively understanding how these drugs exert their effects on the target mRNA. It also sheds light on the efficiency of delivery systems and interactions within the cell. This level of understanding is crucial for making better and safer RNA-based drugs.
- Rapid optimization cycles: Leverage high-content, high-throughput screening capabilities for rapid testing and iteration of design variants to considerably speed up the development of RNA-based drugs.