Sediments from a small mountain lake in Bosnia and Herzegovina preserve an approximately 16,000-year record of summer temperature change in the Western Balkans, revealing abrupt climatic shifts whose effects varied across Europe. Researchers at Eötvös Loránd University reconstructed past mean July temperatures from chironomid remains and other lake-sediment indicators, then used pollen, sediment particles, organic matter and geochemical elements to track associated changes in vegetation, erosion and lake conditions. Two peer-reviewed studies published in 2026 examine different periods in the same sequence, together covering the Late Glacial, the Holocene and more recent centuries.

The reconstructed record shows that a large cooling event associated with the Younger Dryas lowered estimated mean July temperatures at Lake Prokoško by approximately 2.3°C, a change intermediate between estimates from Alpine lakes and those farther east. The lake, which covers 4.8 hectares at an elevation of 1,670 metres in the Vranica Mountains of the Dinaric Alps, also records a later Middle Holocene interval in which estimated summer temperatures may have exceeded present-day levels. These findings emerge from a paleoclimate reconstruction based on biological, chemical and physical evidence preserved in lake sediments, rather than from direct instrumental measurements.

Lake Prokoško occupies a climatically significant position between Western, Central and Eastern Europe. Over thousands of years, remains from the surrounding landscape and the lake’s own ecosystem accumulated in bottom sediments, providing material for reconstructing regional environmental and climatic history. One study, published in Palaeogeography, Palaeoclimatology, Palaeoecology, focused on rapid changes between approximately 16,000 and 10,000 years ago during the Late Glacial and beginning of the Holocene. A separate study in Quaternary Science Reviews traced conditions through the Holocene to the present, extending the interpretation of the sediment sequence across the full period represented by the two investigations.

Chironomids, a group of aquatic insects also known as non-biting midges, supplied the principal temperature indicator. Their larvae occupy lake sediments and differ in their preferences for water temperature, oxygen levels and nutrient conditions. After moulting, larvae leave chitinous head capsules that sink to the lake bed, where they can be preserved for tens of thousands of years. The researchers identified the species represented in different sediment layers and compared the fossil assemblages with faunal and climate data from hundreds of modern European lakes. This comparison supported estimates of mean July temperature, while the accompanying pollen, sediment, organic-matter and chemical records provided evidence about forests, soil erosion, lake productivity and other environmental changes.

The oldest layers examined reflect cold conditions near the end of the last glacial period. At around 14,600 years ago, reconstructed mean July temperatures around Lake Prokoško were below 8°C. They subsequently rose by approximately 2.6°C within a comparatively short interval, reaching around 10.2°C. The reconstruction associates this warming with increased biological activity in the lake, spreading vegetation and more stable conditions in the catchment. The youngest sediments in the study cover the most recent millennium, where the chironomid community reflected not only temperature changes but also centuries of alpine grazing, greater nutrient input and the later expansion of nutrient-rich conditions.

The early warming ended abruptly when the Younger Dryas began around 12,900 years ago, the last major cooling episode at the close of the last glacial period. Estimated mean July temperatures at Lake Prokoško fell by approximately 2.3°C to about 7.9°C. Cold-water chironomid species became more abundant, while increased erosion transported more clastic material into the lake. Comparisons with reconstructions elsewhere in Europe revealed a marked spatial contrast: summer cooling of 3–4°C has been reconstructed at many Alpine lakes, whereas decreases in the Eastern and Southern Carpathians were generally less than 1°C. The Lake Prokoško estimate falls between those geographic extremes.

The researchers connect this west-to-east gradient, which becomes weaker toward the east, to oceanic and atmospheric changes in the North Atlantic region. For the Carpathian Basin, the contrast illustrates that the same large-scale climatic event can produce different responses in Western, Central and Eastern Europe. Lake Prokoško consequently provides a regional reference point for examining how the magnitude of a shared climate signal varied geographically. The researchers also emphasize that alpine lakes respond sensitively to relatively small temperature changes: variations in ice-cover duration, water stratification, oxygen content and biological activity can accompany those changes, making small mountain lakes useful environmental archives.

Warming resumed at the end of the Younger Dryas. Between approximately 11,800 and 11,400 years ago, reconstructed mean July temperatures increased by around 2.9°C. By about 10,500 years ago, they may have reached approximately 12.6°C. The lake became more productive, its ice-free season lengthened, and closed-canopy forests developed in the catchment, reducing erosion. Temperatures did not remain constant throughout the Holocene. Around 6,800 years ago, reconstructed mean July temperatures may have reached approximately 14°C, meaning that some summers around Lake Prokoško were warmer than those experienced today. The researchers distinguish this Middle Holocene warmth from present-day warming: the earlier conditions were mainly seasonal and driven by gradual changes in Earth’s orbital parameters that increased Northern Hemisphere summer insolation, whereas current warming is global, year-round, more rapid and primarily associated with human-caused greenhouse gas emissions.

A cooling trend began around 6,000 to 5,500 years ago during the Neoglacial transition toward a cooler Late Holocene. Reconstructed mean July temperatures fell from approximately 14°C to around 12.8°C, then to 11.8°C by about 5,670 years ago. Pollen evidence indicates that people were altering the landscape around the lake by the end of the Neolithic, approximately 6,000 years ago. Plants associated with grazing and open habitats became more common, followed later by more intensive forest use and the expansion of open areas. The same sediment archive therefore records both reconstructed climatic variability and the growing influence of human land use, although it does not by itself establish that human activity caused the temperature changes documented in the sequence.

Later Holocene temperatures continued to vary. Around 3,000 years ago, another significant cooling began, with mean summer temperatures falling from approximately 14.7°C to 12°C. Around 2,000 years ago, reconstructed temperatures rose during the Roman Warm Period, reaching approximately 13.3°C. Over the past millennium, centuries of alpine grazing opened the surrounding landscape and increased the nutrient load entering Lake Prokoško. The chironomid community changed significantly around 700 years ago, and species preferring more nutrient-rich waters became increasingly common over the following 500 years. Together, the two studies show that the lake’s fossil and sediment record captures environmental responses shaped by both regional climate change and local human transformation, while also demonstrating that estimates for individual periods remain reconstructions derived from sediment proxies and modern comparative data.

Subject of Research: Paleoclimate reconstruction of summer temperature variability in the Western Balkans

Article Title: Where three climate zones meet

Article References: Where three climate zones meet. Original publication (Bibliographic details unavailable.)

Image Credits: AI Generated

DOI: Not provided

Keywords: Lake Prokoško, Western Balkans, summer temperatures, paleoclimate, chironomid fossils, Younger Dryas, Holocene, climate reconstruction, land use, alpine lake, three, climate

Witness the faces of Rome’s forgotten working class.

A neolithic warrior gets a face thanks to modern tech.

In a groundbreaking advance that bridges molecular biology and evolutionary science, researchers at The University of Osaka have unveiled a novel methodology to resurrect ancestral proteins, specifically rhodopsins, enabling direct experimental investigation of these ancient molecules. This work, published in ACS Omega, outlines an innovative sequence reconstruction technique that integrates structural information and attentively accounts for insertions and deletions in protein sequences — a crucial improvement over conventional alignment methods that struggle with highly variable protein domains.

Microbial rhodopsins constitute a diverse family of membrane-embedded proteins with critical biological functions, ranging from ion transport to photoreception. Despite sharing a common structural feature of seven transmembrane helices, these proteins exhibit remarkable functional diversity, largely attributed to extensive variation in their extramembrane domains. Traditional sequence alignment algorithms fall short in tracing their evolutionary trajectories, as they typically do not handle complex insertions and deletions (indels) effectively, especially in variable regions outside of transmembrane segments.

Addressing these challenges, the team led by Haruto Ishikawa and Yasuhisa Mizutani developed ConsistASR, a structure-guided and indel-aware sequence reconstruction pipeline. This approach leverages three-dimensional structural insight to inform sequence alignment, thereby improving the accuracy of reconstructing full-length ancestral rhodopsin proteins. By carefully modeling sequence variability in extramembrane domains, the researchers were able to faithfully infer ancestral sequences that maintain key structural and functional attributes.

Focusing on two subfamilies of microbial rhodopsins—schizorhodopsins and heliorhodopsins—the study successfully resurrected their ancestral versions and expressed these proteins in Escherichia coli. The functional assays revealed that the ancestral schizorhodopsin exhibited light-driven proton pumping activity, consistent with its modern descendants, whereas the ancestral heliorhodopsin lacked ion transport activity, perfectly aligning with the known characteristics of extant heliorhodopsins.

Notably, the ancestral schizorhodopsin protein exhibited the expected spectral properties and stability, signifying correct folding within the bacterial membrane environment. This functional proof of concept not only validates the reconstruction pipeline but also opens a window into the evolutionary dynamics that have shaped rhodopsin diversification over time. The ability to experimentally produce and assay these ancient proteins allows scientists to study evolutionary adaptations at a molecular level with unprecedented precision.

The significance of this work extends beyond just the rhodopsin family. The ConsistASR platform provides a powerful toolkit that can be generalized to reconstruct other ancestral proteins characterized by variable length sequences and complex indels. Such reconstructions are invaluable for evolutionary biology, synthetic biology, and protein engineering, providing empirical insights into how proteins have evolved to acquire new functions and properties.

This innovative integration of sequence, structural, and functional analysis reflects a broader trend in molecular evolution research, where computational biology synergizes with experimental validation to overcome traditional obstacles. As the genomic and structural databases continue to expand, methods like ConsistASR could revolutionize our understanding of protein evolution and functionality by enabling direct interrogation of ancient biomolecules.

The study also tackles a fundamental evolutionary biology question—how functional diversity arises within a protein family while preserving core structural frameworks. By showing that ancestral proteins can be reconstructed in full length and retain distinct biophysical and functional profiles, the research provides direct empirical evidence for evolutionary trajectories that have resulted in contemporary diversity.

Moreover, the expression of these ancestral rhodopsins in bacterial hosts exemplifies how modern synthetic biology tools can resurrect and characterize long-extinct molecular functions. This not only informs evolutionary theory but also offers potential applications in biotechnology, where ancestral proteins might possess unique stability or functional features suitable for industrial or medical purposes.

In practical terms, illuminating the evolutionary path of microbial rhodopsins aids in comprehending microbial adaptation to various ecological niches, given these proteins’ pivotal role in light-driven energy conversion and environmental sensing. Ancestral protein resurrection thus provides a tangible bridge connecting molecular structure, evolution, and ecological function.

The research was supported by the Japan Society for the Promotion of Science and involved meticulous data analysis strategies, combining phylogenetics, structural biology, and bioinformatics. It underscores the multidisciplinary nature of modern biological discovery, illustrating that resolving complex biological problems often requires integrating diverse scientific perspectives and technologies.

Ultimately, this powerful approach reinvigorates the field of ancestral protein research by not only surmounting critical technical obstacles but also by delivering concrete functional data that validate hypotheses about ancient protein behavior. The availability of the ConsistASR workflow to the broader scientific community encourages collaborative exploration of protein evolution and engineering, heralding a new era of experimentally validated molecular paleobiology.

Subject of Research:
Not applicable

Article Title:
Resurrecting Full-length Ancestral Schizorhodopsins and Heliorhodopsins with Structure-guided, Indel-aware Sequence Reconstruction

News Publication Date:
8-Jun-2026

Web References:
https://doi.org/10.1021/acsomega.6c03010

Image Credits:
Haruto Ishikawa

Keywords:
Molecular biology, Biomolecules, Amino acids, Biomolecular structure, Domain organization, Bacterial proteins, Ancient proteins, Membrane proteins, Extracellular domains

In a groundbreaking development that promises to revolutionize the way artificial intelligence perceives the world, a collaborative research team from KAIST, MIT, and Microsoft has unveiled a novel technology named “Upsample Anything.” This pioneering innovation substantially enhances the visual acuity of AI systems while drastically reducing the memory demand on hardware, specifically GPU memory. The new method achieves up to a sixteenfold increase in memory efficiency, addressing one of the most critical bottlenecks in real-time AI vision applications for mobile and embedded devices.

The cornerstone of this advancement lies in overcoming the traditional compromise faced by AI-driven image processing systems. Historically, achieving higher resolution visual input required extensive GPU memory and computational power, creating a significant barrier to real-time implementation on resource-constrained devices such as smartphones and humanoid robots. Conversely, compressing images into low-resolution features, which is a common practice to conserve resources, often leads to the loss of crucial visual details like small objects and fine structural elements. This trade-off has long hindered the deployment of precise AI vision in dynamic and mobile environments.

The “Upsample Anything” technology ingeniously sidesteps this issue by introducing a training-free upsampling mechanism capable of reconstructing high-resolution features from low-resolution inputs without additional data training. Unlike conventional methods that necessitate retraining or complex optimization when exposed to new data or environmental conditions, this approach adapts instantly by leveraging the edge and structural information inherent within a single input image. This capability drastically simplifies deployment and customization across diverse AI vision applications.

At the heart of the method is a sophisticated process called test-time optimization (TTO), which refines pixel-wise anisotropic kernel parameters based on the specific low-resolution image. These learned kernels are then applied to foundational feature maps to generate high-resolution representations. The final phase involves pixel-wise anisotropic Joint Bilateral Upsampling, a technique that preserves edge details and texture fidelity while reconstructing the image in high resolution. This multi-stage pipeline ensures that essential visual information is restored with remarkable accuracy and efficiency.

Highlighted through an illustrative example, the research team demonstrated the restoration of a 224×224 pixel image— a common resolution in AI research— to a clarity nearly identical to the original, within approximately 0.4 seconds of computation. This substantial leap in processing speed and memory efficiency heralds a new era for real-time AI vision, where systems can now maintain high precision without the prohibitive cost of extensive hardware resources. The direct implication is that AI can better understand and interact with its environment, making it invaluable for applications requiring fine object detection and manipulation.

The utility of “Upsample Anything” stretches far beyond humanoid robots. Autonomous driving systems stand to benefit immensely, as they depend heavily on the rapid and accurate interpretation of complex visual scenes. With this technology, vehicles can process crucial visual cues in a more resource-efficient manner, enhancing safety and navigation capabilities. Additionally, on-device AI for smartphones and other portable gadgets can achieve superior visual comprehension while minimizing battery drain and thermal load— critical factors for consumer electronics.

The research team’s achievement was officially recognized at CVPR 2026, the most prestigious conference dedicated to computer vision and artificial intelligence. The work earned the “CVPR Compute Gold Star” award, a distinction granted to research that excels in computational efficiency. Furthermore, it was named the “Transparency Champion” for leading in research process transparency and reproducibility, emphasizing the team’s commitment to responsible and open scientific practices. This dual recognition highlights not only the technological ingenuity but also the ethical and procedural rigor of the research.

One of the most remarkable attributes of this technology is its universality and plug-and-play nature. Its training-free design means that no additional datasets or retraining cycles are required when deploying it in different settings. This contrasts sharply with existing upsampling methods that often demand labor-intensive tuning or environment-specific learning. Consequently, “Upsample Anything” can seamlessly integrate into a wide array of AI frameworks, simplifying engineering workflows and accelerating the timeline from development to deployment.

Professor Changick Kim, the project lead from KAIST’s School of Electrical Engineering, emphasized the transformative potential of this technology. He noted that by boosting visual precision with minimal resource demands, the algorithm can catalyze the commercial viability of complex AI entities like humanoid robots and enable more powerful on-device AI applications. The convergence of computational efficiency and high visual fidelity, he added, represents a critical step forward in the quest to bring sophisticated AI capabilities to everyday devices.

This advancement is particularly significant in the context of emerging AI models based on world models— frameworks that aim to understand and predict changes in physical environments. Effective deployment of such models often requires processing a large volume of visual data accurately and expediently. By enabling high-resolution feature upsampling without overwhelming hardware limitations, “Upsample Anything” paves the way for smarter, faster, and more adaptive AI models that truly comprehend their surroundings in detail.

The first author of the study, Minseok Seo, a doctoral student at KAIST, alongside co-authors from MIT and Microsoft, detailed their findings in a paper titled “Upsample Anything: A Simple and Hard to Beat Baseline for Feature Upsampling.” The work is publicly accessible via the arXiv repository (DOI: 10.48550/arXiv.2511.16301), facilitating further exploration and innovation by the wider research community. Their approach exemplifies a blend of elegance and practicality, challenging the assumption that high computational cost is a prerequisite for high-quality visual reconstruction in AI.

In conclusion, the “Upsample Anything” technology marks a significant leap in AI visual processing by resolving a longstanding dilemma between image resolution and hardware resource constraints. Its immediate applicability across various AI-driven fields— ranging from robotics to autonomous systems and portable devices— underscores its impact and longevity. As AI continues to entwine itself with daily human interaction and technological progress, innovations like this will form the backbone of smarter, more capable, and accessible intelligent systems.

---

Subject of Research: Not applicable

Article Title: Upsample Anything: A Simple and Hard to Beat Baseline for Feature Upsampling

News Publication Date: 17-June-2026

Web References: https://dx.doi.org/10.48550/arXiv.2511.16301

References: Upsample Anything: A Simple and Hard to Beat Baseline for Feature Upsampling, DOI:10.48550/arXiv.2511.16301 (Seo et al., 2026)

Image Credits: KAIST

Keywords

AI vision, upsampling technology, GPU memory efficiency, real-time image processing, test-time optimization, pixel-wise anisotropic kernels, Joint Bilateral Upsampling, on-device AI, humanoid robots, autonomous driving, world models, computer vision, computational efficiency