Liquid Biosciences is a leader in ushering the era of precision medicine by integrating and analyzing a broad set of patient and care data, utilizing proprietary algorithms and multi-omics assays, to better predict clinical outcomes and guide early intervention for customers such as pharmaceutical, biotech companies and healthcare providers.
Out of millions, billions, or trillions of possible solutions, how do you identify what’s most important? Early life on this earth faced this exact problem — out of the billions of possible chemical compounds, nature needed to find the ones that would enable life. A crucial part of this was determining the compounds which would best form a genetic code of all living things. And nature found an incredibly simple and parsimonious solution — at the most simplistic level, four nucleotide base pairs dictate the diversity and complexity of life. Admittedly this isn’t the entire story, but evolution converged on four base pairs that work exceptionally well to form the basis for life on this earth. There is only one proven algorithm that can take millions if and billions of variables and converge on the very most important. This algorithm, evolution, has been vetted through billions of years of the history of the universe.
Precision medicine now faces the same problem that early life did. Healthcare datasets are usually comprised of 50,000 to 2,000,000 variables. Biology, treatment, and behavioral variables are not independent, and interact with each other in ways that are rarely known in advance of analysis. This results in tremendous complexity, dimensionality, and susceptibility to bias from both domain knowledge and mathematical assumptions. Prevalent analytic methods, including the most advanced machine learning, cannot deal directly with the scale of this complexity. These methods also do not include mathematical relationship types that are required to characterize biological/behavioral/treatment systems. For these reasons and others, prevalent analytic methods have demonstrably produced insufficient accuracy and low reproducibility.
We are a bio-analytics company that leverages bioinspired and evolutionary computing techniques to perform multi-omic, multi-modality analyses and drive interpretable and actionable diagnoses, biomarker discovery, and treatment plans. The single biggest challenge of our time is not acquiring data. It’s making sense of it. Its knowing what to focus on. And that is where we excel.