The study was led by Professor Rachel Kolodny and PhD candidate Guy Yanai of the University of Haifa, together with Professor Nir Ben-Tal and graduate student Gabriel Axel of Tel Aviv University, and Specially Appointed Associate Professor Liam M. Longo of ELSI. Kolodny also spent five months as a visiting researcher at ELSI, developing methods to analyze the new model. Their creation, dubbed CLSS for Contrastive Learning Sequence-Structure, is a protein language model designed to overcome a stubborn problem that has limited previous computational approaches: the awkward relationship between what a protein’s sequence says and what its structure actually does.

Scientists have long organized proteins into hierarchical groups based on relatedness, much like the genus and species categories biologists use to classify organisms. These curated systems, such as the widely used ECOD and CATH databases, distill decades of expert knowledge. But with artificial intelligence now capable of generating ’embeddings’—numerical representations in which proteins with similar properties receive nearby coordinates, like postal codes on a map—researchers can visualize relationships across millions of proteins at once, producing what the team calls a protein world map. The catch is that sequence and structure do not map neatly onto each other. Unrelated sequences can fold into similar shapes, while even identical sequences can sometimes adopt wildly different structures.

Most existing protein language models treat sequence and structure as separate worlds, processing one or the other independently. Even hybrid models that incorporate both kinds of data rarely place the sequence and the structure of the same protein at the same location on a global map, leaving researchers with two conflicting atlases of protein space. CLSS was engineered specifically to resolve this discordance. Using a machine learning strategy known as contrastive learning, the model is trained on pairs of protein sequences and their corresponding structures, learning to pull matching sequence-structure pairs together in the embedding space while pushing unrelated pairs apart.

The result is a single shared map in which a protein occupies essentially the same location whether the model is given its sequence or its structure. When benchmarked against other state-of-the-art protein language models, CLSS succeeded in producing a cohesive unified representation, something its predecessors could not achieve. Remarkably, the model’s maps closely reproduced the relationships recorded in the expert-curated ECOD and CATH classification systems, even though those classifications were never shown to the model during training. In direct classification tests, CLSS also performed strongly, demonstrating that merging sequence and structure information yields genuinely more informative protein representations.

Perhaps the most exciting feature of CLSS is its ability to handle fragments. Most protein language models require a complete sequence or structure to generate a meaningful embedding, but CLSS showed that short sequence fragments can in many cases be positioned meaningfully alongside full-length proteins and structures. This capability matters enormously for evolutionary studies, because small pieces of proteins have been repeatedly reused and rearranged throughout the history of life. Some fragments may even have served as the primordial building blocks from which the earliest protein domains were assembled, meaning that similar fragments appearing in otherwise unrelated proteins can hint at ancient evolutionary connections.

The maps produced by CLSS also revealed sweeping patterns across protein space that were previously difficult to see. When the researchers overlaid biological properties onto the maps, proteins associated with organic cofactors turned out to cluster in particular regions, while metal-binding proteins were scattered more broadly. Such patterns illustrate how global protein maps can serve not only as classification tools but as instruments for exploring the interplay between sequence, structure, function, and deep evolutionary history, potentially exposing large-scale patterns invisible to conventional pairwise comparison methods.

‘This gives us a way to look at the protein universe through sequence and structure at the same time, rather than treating them as separate worlds,’ said Longo. ‘What is particularly exciting for us is the possibility of using these maps to uncover large-scale evolutionary patterns that are difficult to recognise using conventional approaches.’ The team ultimately envisions unified sequence-structure representations opening new frontiers in database searches, protein engineering, and the reconstruction of evolutionary trajectories—offering a fresh window onto how the staggering diversity of proteins found in life today emerged over nearly four billion years of evolution.

Subject of Research: A contrastive-learning protein language model that unifies protein sequence and structure representations to map the protein universe

Article Title: Uniting sequence and structure to map the protein universe

Article References: Uniting sequence and structure to map the protein universe. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: protein language model, CLSS, protein evolution, contrastive learning, protein structure, amino acid sequence, ECOD, CATH, protein embeddings, evolutionary biochemistry, protein classification, Institute of Science Tokyo

The emergence of educational AI technologies has posed a significant question: Can machine-generated content effectively evaluate and enhance students’ understanding of intricate scientific principles? Polat and Karadag’s study delves into this very inquiry, experimenting with the parameters of AI-generated biochemistry questions to determine their effectiveness in MST environments. Their findings may not only improve the efficiency of evaluations but could also provide deeper insights into student learning behaviors and comprehension.

This research aims to contextualize the significance of biochemistry in the medical curriculum, emphasizing its role as a foundational subject that connects various facets of medical education. The study’s methodology presents a compelling argument for incorporating AI technologies to generate questions that assess more than just superficial knowledge. By challenging students with diverse, scenarios-based queries, the researchers seek to foster critical thinking and apply theoretical knowledge to practical situations.

One of the key highlights of this study is the deployment of machine learning algorithms to generate a wide range of biochemistry test items. The researchers utilized advanced AI technologies that analyze vast datasets of biochemical information and educational metrics to craft personalized assessments. This automated approach to question generation represents an exciting paradigm shift in educational methodologies, as it allows educators to focus on facilitating knowledge rather than solely crafting evaluations.

Under the MST conditions, the AI-generated questions were designed to challenge students across different levels of understanding. The complexity of the questions varied, simulating a real-world scenario where students must apply their knowledge to solve problems. This method not only assesses knowledge retention but also gauges problem-solving skills, adaptability, and the ability to think on one’s feet—attributes essential for future medical professionals.

Furthermore, the implementation of AI in education carries potential implications beyond just improved assessment tools. It raises ethical considerations regarding the accuracy and bias inherent in machine-generated content. Polat and Karadag acknowledged these challenges by advocating for a collaborative approach that includes continuous human oversight in the AI training process. Such precautions are vital to ensuring the integrity and fairness of the assessments that ultimately shape the future of healthcare professionals.

Additionally, the research investigated the feedback mechanisms in place following MST conditions. After students completed the AI-generated assessments, they were provided with insights into their performance. This not only allowed students to identify knowledge gaps but also enabled them to understand the rationale behind their answers—essential for promoting metacognitive skills. Cultivating self-awareness in learning is critical, as it empowers students to take charge of their educational journeys.

As the study progressed, Polat and Karadag collected data on student performance and perceptions of AI-generated assessments. The responses indicated a general appreciation for the innovative format, with many students expressing that the AI-generated questions were engaging and reflective of real-world applications in biochemistry. This positive feedback is crucial for validating the effectiveness of this new assessment approach.

Moreover, the researchers explored the diverse learning preferences among students and how AI can cater to individualized educational experiences. By analyzing patterns in responses, the AI system could adapt the complexity and style of questions based on student performance, ensuring that every learner’s needs are met. This personalized approach aligns with contemporary educational philosophies focused on learner-centered methods, making education more inclusive and effective.

Notably, the implications of Polat and Karadag’s research extend beyond the classroom. If proven effective, AI-generated assessments could transform standardized testing processes, offering dynamic evaluations that adapt to the test-taker’s knowledge level. This could enhance the overall quality of medical education and create a more robust selection process for future healthcare providers.

Looking ahead, the potential for AI to influence biochemistry education is immense. Future research could expand the range of subjects and disciplines that benefit from such technology, challenging the boundaries of traditional learning environments. Integrating AI into educational frameworks could facilitate innovative approaches to teaching that align with the rapidly advancing scientific landscape.

In conclusion, the study conducted by Polat and Karadag represents a pivotal moment for medical education—a confluence of artificial intelligence and biochemistry that has broad implications for how assessments are designed and delivered. By illuminating the advantages of AI-generated assessment tools, this research may pave the way for transformative changes in educational practices, ultimately leading to a new generation of healthcare professionals equipped for the challenges of modern medicine.

As this research gains traction in the educational community, it will be essential to monitor its implementation and effectiveness through ongoing evaluation and adjustment. The collaboration between AI technologies and human educators will remain paramount, ensuring that AI serves as a complementary tool rather than a replacement for traditional pedagogical methods.

By marrying technology with education, Polat and Karadag’s study not only demonstrates the possibilities that lie ahead but also reinforces the importance of adapting to new realities in the world of learning. The future of biochemistry education, imbued with the ingenuity of AI, promises to be as exciting as it is transformative.

Subject of Research: AI-generated biochemistry test item parameters in MST test conditions

Article Title: AI-generated biochemistry test item parameters in MST test conditions

Article References: Polat, M., Karadag, E. AI-generated biochemistry test item parameters in MST test conditions. BMC Med Educ 25, 1705 (2025). https://doi.org/10.1186/s12909-025-08292-3

Image Credits: AI Generated

DOI: https://doi.org/10.1186/s12909-025-08292-3

Keywords: artificial intelligence, biochemistry education, medical education, MST conditions, AI-generated assessments, personalized learning