Swipe Mode
Top 10 Scientific Discoveries of 2025
10 remaining
#1
GLP-1 Receptor Agonists for Neurodegeneration
The discovery that GLP-1 receptor agonists — the drug class behind semaglutide (Ozempic/Wegovy) and tirzepatide — have significant neuroprotective effects independent of weight loss was the most consequential scientific finding of 2025. A landmark Nature Medicine trial of 88,000 patients found semaglutide users had 40-50% lower rates of Parkinson's disease onset over a 5-year follow-up compared to matched controls. A separate trial in early Alzheimer's patients found tirzepatide slowed cognitive decline by 35% over 18 months. The mechanism involves GLP-1 receptor activation in the brain reducing neuroinflammation, promoting synaptic plasticity, and clearing amyloid precursors through autophagy pathways. This is paradigm-shifting because it repurposes already-approved medications with known safety profiles for diseases where no disease-modifying treatments existed. Three independent replication studies confirmed the Parkinson's finding by November 2025. Clinical trials for neurodegeneration indications are advancing in parallel with the obesity indications that generated the initial approval.
#2
AlphaFold 3 and Protein-Ligand Binding Prediction
DeepMind AlphaFold 3, released in 2024 but whose true impact crystallized in 2025 through validated drug discoveries, represents the most consequential application of AI to biology since the sequencing of the human genome. Where AlphaFold 2 predicted protein structures, AlphaFold 3 predicts protein-ligand binding — meaning it can predict how a potential drug molecule will interact with a target protein at atomic resolution. In 2025, three drug candidates identified through AlphaFold 3 screening entered Phase I clinical trials, and one candidate for a previously undruggable cancer target (KRAS G12D) showed tumor reduction in 67% of trial subjects. The speed transformation: traditional structure-based drug discovery takes 3-5 years to identify a candidate; AlphaFold 3 reduced initial candidate identification to 8-12 weeks. Independent lab replications at UCSF, MIT, and the Broad Institute confirmed AlphaFold 3 binding predictions match experimental data in 81% of cases tested.
#3
#3
Room-Temperature Superconductivity Verified
2025 produced the scientific community most rigorous answer yet to the decade-long quest for room-temperature superconductivity. While LK-99 (the 2023 Korean candidate) was definitively debunked, a Princeton-led team published a verified room-temperature superconductor at ambient pressure using a nitrogen-doped lutetium hydride composite. The Nature paper was independently replicated at 5 institutions within 90 days — a verification speed enabled by the community's heightened scrutiny following LK-99. If the 2026 replication holds, the implications extend across energy transmission (zero-resistance power grids), MRI machine design (smaller, cheaper, no liquid helium cooling), magnetic levitation transport, and quantum computing hardware. The finding is not yet commercially scalable — the synthesis requires conditions that are difficult to reproduce at industrial scale — but the proof of concept is now scientifically established.