Abzu
Abzu AI is a trustworthy and rational AI tool designed to assist in mission-critical decision-making processes. The platform offers rationa...
Last verified:
What is Abzu?
Abzu is a biotechnology company that applies proprietary explainable AI to develop best-in-class RNA therapies for major medical needs. Their core technology is the QLattice®, a supervised machine learning method for symbolic regression that unlocks discoveries in data by revealing insights that fuel innovation. The company combines deep RNA biology, AI-driven design, and drug development experience into a single execution engine for developing better and safer RNA drugs.
The QLattice is designed for modern researchers working in pharma R&D, biotech, and research institutions. It generates explainable models where predictions are simple mathematical equations that are readily interpretable. The tool is small-data friendly, delivering unparalleled performance even on small or wide datasets using as few inputs as necessary. It serves as a discovery engine where models act as hypotheses, revealing complex patterns in data for further analysis.
Abzu's AI-guided design platform allows scientists to explore and prioritize vast sequence space in silico, routinely evaluating 100,000+ design variants and predicting efficacy and key developability properties before entering the lab. This reduces experimental cycles, lowers cost, and shortens time to candidate selection. The platform has been validated for designing active and safe ASO and siRNA therapeutics, and has increased safe ASOs by 20% over in-house models.
The company serves pharmaceutical companies, biotech firms, and research institutions working on RNA therapeutics including siRNA, ASOs, and anti-miRs. Their approach addresses extrahepatic delivery challenges by developing targeted aptamers that act as programmable carriers for cell-specific uptake. The QLattice is ISO 27001 compliant for trustworthy AI and has been recognized by Gartner as a
Abzu pricing
Pricing model: Free
The website does not disclose specific pricing details for the QLattice or Abzu's services. No free tier is explicitly mentioned on the main website. The company appears to work through strategic partnerships with pharma, biotech, and research institutions rather than offering self-service plans. Contact appears to be required for enterprise pricing and partnership discussions.
Abzu pros
- Proprietary explainable AI with transparent mathematical equations
- ISO 27001 compliant for trustworthy AI and security
- Outperforms deep learning models in accuracy
- Superior interpretability compared to linear models and decision trees
- Excellent performance on small or wide datasets
- Evaluates 100,000+ design variants in silico before lab work
- Reduces experimental cycles and lowers development costs
- Shortens time to candidate selection for drug discovery
- Increases safe ASOs by 20% over in-house models
- Gartner Cool Vendor recognized for explainable AI
- Patented technology created and maintained in-house
- Models serve as hypotheses for scientific discovery
- Closed learning loop where data improves models and models improve molecules
- Best-in-class for designing safe ASOs and siRNA therapeutics
- Cookieless platform with easy-to-read privacy policy
Abzu cons
- Focused primarily on pharma R&D and RNA therapeutics, limiting general use
- No publicly available free tier mentioned on website
- Enterprise pricing not disclosed publicly
- Requires Python knowledge to use Feyn documentation
- Small team of only 9 Abzoids may limit support capacity
- Offices only in Copenhagen and Barcelona, limited global presence
- Specialized for symbolic regression, not all ML tasks
- Primarily serves biotech/pharma industry rather than general data science
Frequently asked questions about Abzu
What is the QLattice?
The QLattice is Abzu's proprietary explainable AI and supervised machine learning method for symbolic regression. It was created to challenge black-box AI assumptions and generates predictions as simple mathematical equations that are readily interpretable and explainable. The technology is built and maintained in-house on patents owned by Abzu.
Who is Abzu's AI for?
Abzu's QLattice is designed for pharma R&D teams, biotechnology companies, and research institutions working on RNA therapeutics. It serves scientists developing siRNA, ASOs, and anti-miRs, as well as those working on drug delivery systems including LNP formulations.
How does Abzu's AI compare to deep learning?
The QLattice provides better accuracy than deep learning models while outperforming the interpretability of linear models and decision trees. Unlike deep learning's black boxes, the QLattice delivers clear, explainable insights without compromising performance.
What makes Abzu's AI explainable?
The QLattice generates models where predictions are simple mathematical equations that are readily interpretable. Only explainable AI provides 'the why' behind predictions, giving researchers decisions they can understand rather than opaque outputs.
Can the QLattice work with small datasets?
Yes, the QLattice is specifically designed to be small-data friendly. Performance is unparalleled even on small or wide data sets, using as few inputs necessary. Models developed by the QLattice have unparalleled accuracy even with very little data.
What RNA therapeutics can Abzu design?
Abzu has designed and validated large numbers of siRNAs, ASOs (antisense oligonucleotides), and anti-miRs across multiple partnered programs. Their in-house activity models are best-in-class for designing safe ASOs.
How does Abzu address drug delivery challenges?
Abzu develops targeted aptamers that act as programmable carriers, enabling cell-specific uptake of therapeutic RNA. This modular platform enables precision delivery beyond the liver, addressing the central challenge of extrahepatic delivery for RNA therapeutics.
Is the QLattice ISO certified?
Yes, the QLattice is ISO 27001 compliant for trustworthy AI, ensuring security and trust in collaborative environments. This certification supports the company's commitment to transparency and integrity.
How do I try the QLattice?
You can try the QLattice by installing the Feyn Python module using 'pip install feyn'. The documentation provides quick start guides with complete examples for classification and regression tasks, including a 'Feyn in 5 minutes' learning path.
What awards has Abzu won?
Abzu won the 2023 'Best HealthTech Nordics Award' at the Nordic Startup Awards, as well as winner of both the 'Synthetic Track' and 'Real-world Track'. The company was also recognized by Gartner as a 'Cool Vendor' in AI Governance and Responsible AI.