At Koliber Biosciences we are passionate about utilizing data science to accelerate biological discoveries. We offer a full range of solutions from traditional statistical analyses and data visualization to innovative machine learning and artificial intelligence systems.
Our mission is to help researchers and business leaders across a variety of biotech and healthcare sectors to incorporate powerful predictive algorithms into their everyday research and turn complex data into key insights.
Machine Learning and Artificial Intelligence solutions
Biological data is inherently complex, expensive to generate and often prone to poor reproducibility. Leveraging innovative tools such as machine learning can help overcome the challenges with utilizing biological data to make progress on a research project. Machine learning can be utilized to build predictive models that enable pre-screening of novel variants or conditions before testing in the wet lab. Machine learning can also identify key variables that drive performance and generate novel insights. Application areas include: drug discovery, diagnostics, process / assay optimization, genome engineering and more.
At Koliber we build custom machine learning models from biological data. We also help customers identify what and how much data to generate to achieve specific performance. We have significant domain expertise in a variety of techniques including: feature based machine learning, deep learning, data augmentation, transfer learning and explainable AI.
Ready to try machine learning but lack the team or capacity to try it? Contact us to discuss a feasibility project.
Available today.
AI platform for peptide drug discovery
Koliber Biosciences has developed an AI assisted peptide drug discovery platform. The platform is ideal for analysis of complex datasets derived from peptide variants that include multiple performance metrics such as activity, selectivity, solubility, and in vivo results. Two classes of encodings are available: classical peptide features such as physiochemical properties and proprietary deep learning encodings. The encodings are combined with tunable algorithms and model architectures that are suitable for datasets generated from as few as 100 peptides. The result is a system that allows the user to: 1) pinpoint the features that drive performance; 2) determine which in vitro assays predict in vivo results; 3) identify sites that are important for activity; and 4) select new variants that are most likely to have improved activity. The platform has been validated on several peptide datasets, has been funded by the National Science Foundation and is currently available for early access.
Statistics Training and Consulting
Statistics can be overwhelming for researchers trained in biological sciences. The learning curve is steep and the training materials full of old jargon and irrelevant examples. Yet in today’s data heavy world, statistical analysis is essential to make informed decisions on which research avenues to pursue. At Koliber we flatten the learning curve by providing hands on training on datasets generated by the customer thus focusing on the relevant learning while providing an immediate benefit. The training utilizes commercial software that is easy to use and can be adopted by scientists at every level of the organization from lab associates to senior leaders.
Koliber also specializes in advanced statistical methods such as Design of Experiments and Six Sigma, methods that can be utilized for guiding efficient multi factorial experimental design and for identifying parameters that lead to robust and reproducible process or assay performance.
Ready to implement the correct statistical analyses at your organization? Contact us to discuss training or custom statistical solutions.
Microbiome and Probiotics Engineering
Koliber Biosciences is developing genome engineering technologies for microbiome and probiotic strains. The technology platform will enable the development of novel microbiome therapeutics, accelerate microbiome research and provide the ability to add novel functions into existing probiotic strains. The project has been funded by the National Institutes of Health NIGMS SBIR grant and resulted in a suite of vectors for engineering Lactobacillus species. We leveraged machine learning approaches to develop novel genetic control elements that enable robust performance across a broad range of species. We also developed robust methods for DNA delivery and demonstrated genome editing in several species. Contact us to learn more about early access to the genetic engineering tools.
12265 World Trade Drive, Suite G, San Diego, CA 92128, USA
(858) 521-8484
info@koliberbio.com
koliber biosciences news
june2020
SBIR phase 1 award - Artificial Intelligence Platform for Peptide Drug Discovery
may2020
World Trade Drive - Koliber Biosciences moves to expanded facilities - location
april2017
SBIR phase 1 award - Genetic toolkit for genome editing of the Lactobacillus genus.
november2015
SynBioBeta SF2015 - Ewa Lis, CTO presents at Syn Bio Beta in San Francisco, describes how machine learning can be used to accelerate commercialization of synthetic biology products. SynBiobeta SF 2015
- Koliber Biosciences wins BlueSkyBio Competition
- Koliber Biosciences receives $51,000 synthetic DNA credits from Twist
january2015
SBIR phase 1 award - Koliber Biosciences awarded a phase I SBIR grant from the National Science Foundation for the development of microbial capsaicin production platform. NSF Award
december2014
Bio Tech and Beyond - Koliber Biosciences joins the Bio Tech and Beyond incubator in Carlsbad, CA.
Made with Mobirise