Transforming RNA therapeutics with cutting-edge AI.
Directly from target to in vitro validated candidates.
Expertise.
The Abzu team has deep experience in designing successful siRNAs, anti-miRs, and ASOs.
Data-driven drug design process.
Better designs with AI predicting important drug properties.
We use our AI-powered drug design suite to create designs, predict drug properties, and detect potential off-targets.
Design viable compounds for a target.
Our optimized and smart tiling library allows us to design viable compounds for a given target, in connection with the rest of our RNA design pipeline to enrich the designs and predict which compounds are most likely to succeed.
Use our pre-trained models to predict carefully selected drug properties.
We have an ever-expanding suite of drug property prediction models that have been trained and validated on data our data scientists have curated and used successfully in projects. This allows us to choose the predictive models that are most relevant for your specific use case.
Identify the compounds likely to fail on one or more parameters.
Our design pipeline provides fine-grained control over the entire selection process, and allows establishing and customizing additional criteria for when to reject a compound based on the outcome of our model predictions.
Fine-tune and update the candidate selection process.
In addition to our in silico screening, we can design and perform targeted high quality experiments to fine-tune and update the candidate selection process and achieve better results.
Ready-to-use compound designs.
The outcome is a list of ready-to-use compound designs and a report that outlines the selection process, including how our models have weeded out compounds that are likely to fail on important metrics.
Drug property modelling.
Because successful drug design is about more than activity.
Through years of successful drug design, we’ve continued to improve and expand our selection of models for drug properties. This development stands on the shoulders of our own powerful and explainable AI technology.
Predictors.
Below are some examples of models that we’re actively using to deliver better designs.
Potential off-targets.
Activity.
Stability.
Toxicity.
Species cross-reactivity.
Genetic variations.
Thermodynamics.
Physicochemical properties.
We have an in-house RNA design pipeline connecting everything in one package.
It is built from bespoke functionality that enables us to create and validate data at a rapid speed, including our tailor-made fast oligo sequence aligner that allows for computation in hours rather than days.
Data and R&D.
High quality data for high quality predictions.
Our models are continuously improved and validated on data that we generate ourselves, as well as data that is found in the public – such as from patents.
We maintain an in-house data pipeline that allows our team of dedicated people to curate and process this data to make it suitable for machine learning and validation of our models.
We also have a number of research projects ongoing at any given moment to further expand our offering and apply to different areas.
If you have something specific in mind, reach out and we’ll talk about how we can help.
Custom models.
Add your own research to the mix.
Every project is different, which is why we also offer working with you to featurize* your existing screening data in order to train custom models suited for exactly your needs.
We combine this with our existing suite of models to achieve the best possible predictions. This allows you to leverage the research you’re already doing and fast-track your next drug design project.
*Featurization is the process of taking raw data and ensuring it is in suitable condition for machine learning and AI purposes. Not all data will be of sufficient quality, but we have the expertise to help you maximise what you have and how to navigate from there.
Abzu as a one-stop solution from design to in-vitro validation.
To supercharge the drug design process, we offer ourselves as a one-stop shop for design and in vitro screening for those who want the full benefits of data-driven drug design.
Reduced project management overhead.
Reduced design time.
Directly from target to in vitro validated candidates.
Scientific approach to data.
We design experiments to a high standard that ensures the necessary quality for AI and machine learning, which wouldn’t be obtained otherwise. This method maximises learning by grounding our models in real-world results and reduces the overall time-to-result in a reliable way.
We do our own data generation, research and modelling. This means you leverage not only your existing data, but also all of our data and models we’ve trained from the get go.
Case studies.
Subscribe for
notifications from Abzu.
