HSE Scientists Develop DeepGQ: AI-based 'Google Maps' for G-Quadruplexes
Researchers at the HSE AI Research Centre have developed an AI model that opens up new possibilities for the diagnosis and treatment of serious diseases, including brain cancer and neurodegenerative disorders. Using artificial intelligence, the team studied G-quadruplexes—structures that play a crucial role in cellular function and in the development of organs and tissues. The findings have been published in Scientific Reports.
DNA can be thought of as a long chain of symbols made up of four letters: A, C, G, and T. Beyond the sequence itself, the way the DNA strand is folded and twisted is also crucial. This structure determines which regions of the genetic code are open—accessible to the cell for reading and replication—and which remain closed. One form of such 'packaging' is a special structure known as a G-quadruplex. It can be visualised as a small nodular structure that forms in regions rich in guanine (G). Scientists have hypothesised that each cell type has its own unique set of such nodules, which helps determine the cell’s function. For example, the DNA of nerve cells in the brain differs from that of liver cells in the pattern of these specialised structures. These variations influence how different cell types develop and function. However, investigating these processes in the laboratory is expensive and does not always produce reliable results.
Researchers at the HSE AI Research Centre have developed the DeepGQ AI model, which uses deep learning to generate tissue-specific maps of G-quadruplexes. The model analyses the DNA strand in both directions simultaneously, much like reading it from left to right and from right to left. This bidirectional approach allows the program to capture a complete and accurate picture of the features of the DNA region under study.
Artem Bashkatov, Junior Research Fellow at the Centre for Biomedical Research and Technology of the AI and Digital Science Institute, explains: 'We hypothesised that the characteristics of the cellular environment—not only the DNA structure itself—play a key role in determining how different tissue-specific cell types develop. To test this idea, we developed not a single universal model, but a set of specialised DeepGQ models tailored to different tissues. For example, one model was trained exclusively on brain cells (DeepGQ Neurons), another on liver cells (DeepGQ Liver), and so on across 14 tissue types. This approach enabled each model to identify developmental patterns specific to its tissue.'
Thanks to DeepGQ, scientists now have access to a tool that enables highly accurate prediction of G-quadruplexes. Rather than relying on costly experimental studies, any laboratory investigating, for example, liver cancer or Alzheimer’s disease can use data from patient samples and apply DeepGQ to generate an accurate map of the most likely targets for experimental validation.
'DeepGQ is essentially like Google Maps for G-quadruplexes: it plots landmarks (GQs) and traffic conditions (DHSs and histones) onto a flat DNA map that is unique to each "city," or tissue,' says Maria Poptsova, Director of the Centre for Biomedical Research and Technology at the Institute of AI and Digital Sciences. 'Many serious diseases—from cancer to neurodegeneration—are disorders of lost tissue identity. Cells either forget what they are meant to become, or their specialisation programmes break down. G-quadruplexes may become promising new targets for the treatment of such diseases. If a cancer cell depends on a specific GQ structure for its survival, it may be possible to develop a drug that either disrupts or stabilises this structure until the cellular programme collapses, ultimately killing the cancer cell. In the future, this approach could lead to the creation of a DeepGQ Patient model: by analysing data from a tumour biopsy, researchers could generate a personalised map of active G-quadruplexes and use it to design a truly individualised treatment strategy.'
See also:
HSE Researchers Train Neural Network to Predict Protein–Protein Interactions More Accurately
Scientists at the AI and Digital Science Institute of the HSE Faculty of Computer Science have developed a model capable of predicting protein–protein interactions with 95% accuracy. GSMFormer-PPI integrates three types of protein data (including information about protein surface properties) to analyse relationships between proteins, rather than simply combining datasets as in previous models. The solution could accelerate the discovery of disease molecular mechanisms, biomarkers, and potential therapeutic targets. The paper has been published in Scientific Reports.
HSE MIEM and MTS Launch Workshop on Innovative Solutions in Communication Networks
The HSE Tikhonov Moscow Institute of Electronics and Mathematics (MIEM) and MTS are launching a joint workshop in which students will work at the intersection of communications network engineering, data analysis, and digital technologies. The project is designed as a practice-oriented learning format, enabling students to tackle real industry challenges alongside company engineers and MIEM specialists. Registration to participate in the workshop is open until April 15, 2026.
HSE Scientists Uncover Mechanism Behind Placental Lipid Metabolism Disorders in Preeclampsia
Scientists at HSE University have discovered that in preeclampsia—one of the most severe complications of pregnancy—the placenta remodels its lipid metabolism, reducing its own cholesterol synthesis while increasing cholesterol transfer to the foetus. This compensatory mechanism helps sustain foetal nutrition but accelerates placental deterioration and may lead to preterm birth. The study findings have been published in Frontiers in Molecular Biosciences.
HSE Experts Reveal Low Accuracy of Technology Forecasts in Transportation
HSE researchers evaluated the accuracy of technology forecasts in the transportation sector over the past 50 years and found that the average accuracy rate does not exceed 25%, with the lowest accuracy observed in aviation and rail transport. According to the scientists, this is due to limitations of the forecasting method and the inherent complexities of the sector. The study findings have been published in Technological Forecasting and Social Change.
Wearable Device Data and Saliva Biomarkers Help Assess Stress Resilience
A team of scientists, including researchers from HSE University, has proposed a method for assessing stress resilience using physiological markers derived from wearable devices and saliva samples. The participants who adapted better to stress showed higher heart rate variability, higher zinc concentrations in saliva, and lower potassium levels. The findings were published in the Journal of Molecular Neuroscience.
HSE Unveils Anthropomorphic Courier Robot
From April 1 to 3, 2026, the Fourth Robotics Festival took place, with the HSE Faculty of Computer Science acting as the main organiser. The event featured the presentation of the anthropomorphic courier robot Arkus. The humanoid was introduced by the Institute for Robotic Systems, established jointly by HSE University and the EFKO Group of Companies.
When Circumstances Are Stronger Than Habits: How Financial Stress Affects Smoking Cessation
HSE researchers have found that the likelihood of quitting smoking rises with increasing financial struggles. While low levels of financial difficulties do not affect smoking behaviour, moderate financial stress can increase the probability of quitting by 13% to 21%. Responses to high financial stress differ by gender: men are almost 1.5 times more likely to give up cigarettes than under normal conditions, whereas no significant effect is observed on women’s decisions to quit smoking. These conclusions are based on data from the Russia Longitudinal Monitoring Survey (RLMS-HSE) for 2000–2023 and have been published in Monitoring of Public Opinion: Economic and Social Changes.
HSE Researchers Propose New Method of Verbal Fluency Analysis for Early Detection of Cognitive Impairment
Researchers from the HSE Center for Language and Brain and the Mental Health Research Centre have proposed a new method of linguistic analysis that enables the distinction between normal and pathological ageing. Using this approach, they showed that patterns in patients’ word choices during verbal fluency tests allow clinicians to more accurately differentiate clinically significant impairments from subjective memory complaints. Incorporating this type of analysis into clinical practice could improve the accuracy of early dementia diagnosis. The results have been published in Applied Neuropsychology: Adult.
How the Brain Processes a Word: HSE Researchers Compare Reading Routes in Adults and Children
Researchers from the HSE Center for Language and Brain used magnetoencephalography to study how the brains of adults and children respond to words during reading. They showed that in children the brain takes longer to process words that are frequently used in everyday speech, while rare words and pseudowords are processed in the same way—slowly and in parts. With age, the system is reorganised: high-frequency words shift to a fast route, whereas new letter combinations are still analysed slowly. The study was published in the journal Psychophysiology.
How Neural Networks Detect and Interpret Wordplay: New Insights from HSE Researchers
An international team including researchers from the HSE Faculty of Computer Science has presented KoWit-24, an annotated dataset of 2,700 Russian-language Kommersant news headlines containing wordplay. The dataset enables an assessment of how artificial intelligence detects and interprets wordplay. Experiments with five large language models show that even advanced systems still make mistakes, and that interpreting wordplay is more challenging for them than detecting it. The results were presented at the RANLP conference; the paper is available on Arxiv.org, and the dataset and the code for reproducing the experiments are available on GitHub.