CAS Insights^TM

Patent search looks like a retrieval problem, which is why general-purpose AI struggles with it. In a recent webinar, patent searchers and AI engineers explained where models fail when reading patent text, and what must sit underneath AI before its answers can be trusted.

Carbon dioxide is a greenhouse gas, but in its supercritical state it functions differently: as a non-toxic solvent that is biocompatible for use in pharmaceuticals, power generation, and more. Industrial applications are growing and are driving carbon sequestration efforts.

Genome AI language models wrote complete bacteriophage genomes from scratch, and 16 of the synthesized designs became viruses with antibacterial activity. A cocktail of those phages cleared E. coli strains that were resistant to the natural template phage.

Nanozymes are materials at the nanoscale that act like natural enzymes, but they’re more stable and can be tuned to have specific properties. While largely still in the research phase, these tiny catalysts could enable major breakthroughs in diagnostics, therapeutics, sensors, and environmental remediation.

Webinar recap: A patent-based analysis of where pharmaceutical innovation is heading, from diversifying therapeutic modalities and expanding molecular targets to the disease areas where research is accelerating.

Polymer science now spans millions of possible sequences and hundreds of billions in market value, making it harder than ever to spot what matters. A new joint report with PetroChina SAMRI maps the emerging topics shaping energy storage, sustainable materials, flexible electronics, and membranes.

Classical accuracy metrics assume one correct answer per question. Language models (LLMs) and agentic AI break that assumption. Learn about evaluation approaches emerging for scientific AI.

Deep eutectic solvents are key innovations for green chemistry. They’re cost-effective, environmentally friendly substances that can be used in drug delivery, wastewater treatment, and environmental clean-up. Researchers are finding new substances that make them more effective.

A first-of-its-kind data analysis of 25 years of published process chemistry reactions, conducted by CAS and Lilly, reveals long-term trends in solvents, catalysts, reaction conditions, and emerging methods at an industrial scale.

Immunotherapy is driving precision oncology, but existing immuno-oncology treatments aren’t effective for every patient. Our review of the translational research landscape reveals which targets and therapies are showing promising results in clinical trials.

AI tools are vital to patent searching, but they must have access to structured, indexed data to return reliable, comprehensive results.

2D materials like graphene and carbon nanotubes are revolutionizing materials science, and now MXenes are introducing even greater adaptability and high-performance properties. Find out why MXenes are versatile industrial materials.

Learn how composting has evolved into a precision-engineered bioprocess, and about how aeration, microbial control, materials science, and AI are reshaping the field.

Five years of global pharmaceutical patent data reveal a field in fundamental transition: therapeutic platforms multiplying, molecular targets expanding well beyond traditional proteins, and disease focus shifting toward previously intractable conditions. The patterns show not just where innovation

Aerogels in shin guards, self-powered sensors in jerseys, leucine-tuned recovery protocols, and smarter artificial turf. The science behind today's most-watched sports is moving faster than the players.

As a new Ebola outbreak emerges, research into this devastating disease continues to accelerate. Discover how scientists are developing vaccines and therapeutics while confronting the challenge of sustaining innovation between crises.

Three experts from CAS, the University of Messina, and the Coriell Institute break down how epigenetics has evolved from a niche research area into a clinical force, with 13 FDA-approved drugs, 2,200+ active trials, and a fast-growing diagnostic toolkit reshaping how we detect and treat cancer.

Highlights from a recent CAS Insights webinar on AI in medicinal chemistry. How CAS BioFinder was used to assist in target prediction for bioactive compounds.

For high-performance technologies in power electronics, energy storage, and optoelectronics, semiconductors need to offer performance beyond standard silicon. Wide bandgap semiconductors can handle the high‑voltage, high‑temperature, and high‑frequency conditions of these emerging applications.

Hantavirus, a rodent-borne RNA virus with case fatality rates up to 40%, is reshaping zoonotic disease risk amid environmental change and global mobility.

Quantum computing includes many methods and types of qubits, and certain materials are key for developing these computing capabilities.

Cyclic peptides are improving treatments for cancer, inflammatory and infectious diseases, and more as oral delivery methods are improved.

Scientific R&D teams face real barriers to AI adoption. Learn how structured data, expert collaboration, and the Triangle of Success framework drive meaningful AI maturity.

Vector-borne diseases are spreading to new regions as climate change intensifies, and public health systems need more treatment options to limit the impacts.

Spiro[2.3]hexanes are an example of promising yet challenging scaffold structures for new drug compounds. By using AI-powered predictive analytics with CAS BioFinder, researchers more efficiently identified candidates for validation.