Zaikinlab on Systems Medicine: Dealing with Complexity
You are welcome to the webpage of Systems Medicine Group! This is a unique place, where you, as a PhD or master student,
can both write mathematical equations to model medical or biological data and take part in real wet biological experiments.
Here you can find general information about our approach. For more details about our current projects go to " Research". Systems Medicine is part of Systems Sciences, it means that we are researching general, system-independent and universal laws which govern the behaviour of complex systems. We apply this approach to medical data, spanning cancer diagnostics, prostate cancer active surveillance, obstetric outcomes prediction, COVID-19 and post-COVID modelling, and neuroscience. To perform effective analysis of medical data we develop new statistical and AI-based methods, including Bayesian statistics, Markov Chain Monte Carlo methods, deep learning (LSTM, graph neural networks, reservoir computing), explainable AI, and our novel synolitic network approach. At the same time, to verify our theoretical research on modelling complex biological systems we collaborate with wet biologists and clinicians worldwide. Schematically, our approach can be represented with the following diagram.
Our research now encompasses twelve major directions: from fundamental noise-induced phenomena and nonlinear dynamics, through modelling of proteasomal function and bone dynamics, to applied cancer diagnostics (ovarian, pancreatic, lung, breast, colorectal, and prostate cancers), DNA methylation analysis, AI-driven data analysis using parenclitic and synolitic networks, obstetric prediction models, and computational neuroscience including neuron-astrocyte networks, consciousness modelling, brain ageing, Alzheimer's disease, and sleep cycle dynamics.
A central theme is the development of Synolitic Graph Neural Networks (SGNNs), which have demonstrated remarkable performance in brain imaging classification and cancer detection, overcoming the curse of dimensionality in high-dimensional biomedical data. We have also contributed to the emerging field of bio-artificial intelligence, from genetic perceptrons to emergent digital bio-computation in bacterial populations.
More specifically, if we discuss application of Systems approach to Medicine, the main question is finding the association between all possible -omes, which we can measure with the disease:
