KZKsenia Zlobina

Ksenia Zlobina

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Ksenia Zlobina has a total of 19 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Accelerating Wound Healing Through Deep Reinforcement Learning: A Data-Driven Approach to Optimal Treatment

    by Fan Lu, Ksenia Zlobina, Prabhat Baniya, Houpu Li, Nicholas Rondoni, Narges Asefifeyzabadi, Wan Shen Hee, Maryam Tebyani, Kaelan Schorger, Celeste Franco, et al.

    Advancements in bioelectronic sensors and actuators have paved the way for real-time monitoring and control of the progression of wound healing. Real-time monitoring allows for precise adjustment of treatment strategies to align them with an individual’s unique biological response. However, due to the complexities of human–drug interactions and a lack of predictive models, it is challenging to determine how one should adjust drug dosage to achieve the desired biological response. This work proposes an adaptive closed-loop control framework that integrates deep learning, optimal control, and reinforcement learning to update treatment strategies in real time, with the goal of accelerating wound closure. The proposed approach eliminates the need for mathematical modeling of complex nonlinear wound-healing dynamics. We demonstrate the convergence of the controller via an in silico experimental setup, where the proposed approach successfully accelerated the wound-healing process by 17.71%. Finally, we share the experimental setup and results of an in vivo implementation to highlight the translational potential of our work. Our data-driven model suggests that the treatment strategy, as determined by our deep reinforcement learning algorithm, results in an accelerated onset of inflammation and subsequent transition to proliferation in a porcine wound model.


  • Open Access

    Enhancing wound healing through deep reinforcement learning for optimal therapeutics

    by Fan Lu, Ksenia Zlobina, Nicholas A. Rondoni, Sam Teymoori, Marcella Gomez

    Finding the optimal treatment strategy to accelerate wound healing is of utmost importance, but it presents a formidable challenge owing to the intrinsic nonlinear nature of the process. We propose an adaptive closed-loop control framework that incorporates deep learning, optimal control and reinforcement learning to accelerate wound healing. By adaptively learning a linear representation of nonlinear wound healing dynamics using deep learning and interactively training a deep reinforcement learning agent for tracking the optimal signal derived from this representation without the need for intricate mathematical modelling, our approach has not only successfully reduced the wound healing time by 45.56% compared to the one without any treatment, but also demonstrates the advantages of offering a safer and more economical treatment strategy. The proposed methodology showcases a significant potential for expediting wound healing by effectively integrating perception, predictive modelling and optimal adaptive control, eliminating the need for intricate mathematical models.


  • Open Access

    Robust classification of wound healing stages in both mice and humans for acute and burn wounds based on transcriptomic data

    by Ksenia Zlobina, Eric Malekos, Han Chen, Marcella Gomez

    Background: Wound healing involves careful coordination among various cell types carrying out unique or even multifaceted functions. The abstraction of this complex dynamic process into four primary wound stages is essential to the study of wound care for timing treatment and tracking wound progression. For example, a treatment that may promote healing in the inflammatory stage may prove detrimental in the proliferative stage. Additionally, the time scale of individual responses varies widely across and within the same species. Therefore, a robust method to assess wound stages can help advance translational work from animals to humans.Results: In this work, we present a data-driven model that robustly identifies the dominant wound healing stage using transcriptomic data from biopsies gathered from mouse and human wounds, both burn and surgical. A training dataset composed of publicly available transcriptomic arrays is used to derive 58 shared genes that are commonly differentially expressed. They are divided into 5 clusters based on temporal gene expression dynamics. The clusters represent a 5-dimensional parametric space containing the wound healing trajectory. We then create a mathematical classification algorithm in the 5-dimensional space and demonstrate that it can distinguish between the four stages of wound healing: hemostasis, inflammation, proliferation, and remodeling.Conclusions: In this work, we present an algorithm for wound stage detection based on gene expression. This work suggests that there are universal characteristics of gene expression in wound healing stages despite the seeming disparities across species and wounds. Our algorithm performs well for human and mouse wounds of both burn and surgical types. The algorithm has the potential to serve as a diagnostic tool that can advance precision wound care by providing a way of tracking wound healing progression with more accuracy and finer temporal resolution compared to visual indicators. This increases the potential for preventive action.


  • Open Access

    Effective Spatio-Temporal Regimes for Wound Treatment by Way of Macrophage Polarization: A Mathematical Model

    by Ksenia Zlobina, Jiahao Xue, Marcella Gomez

    Wound healing consists of a sequence of biological processes often grouped into different stages. Interventions applied to accelerate normal wound healing must take into consideration timing with respect to wound healing stages in order to maximize treatment effectiveness. Macrophage polarization from M1 to M2 represents a transition from the inflammatory to the proliferative stage of wound healing. Accelerating this transition may be an effective way to accelerate wound healing; however, it must be induced at the appropriate time. We search for an optimal spatio-temporal regime to apply wound healing treatment in a mathematical model of wound healing. In this work we show that to maximize effectiveness, treatment must not be applied too early or too late with respect to peak inflammation. We also show that the effective spatial distribution of treatment depends on the heterogeneity of the wound surface. In conclusion, this research provides a possible optimal regime of therapy that focuses on macrophage activity and a hypothesis of treatment outcome to be tested in future experiments. Finding optimal regimes for treatment application is a first step toward the development of intelligent algorithms for wound treatment that minimize healing time.


  • Open Access

    Transcriptomic Time Series Analysis in Wound Healing: Challenges and Perspectives on Data Interpretation

    by Ksenia Zlobina, Eric Malekos, Han Chen, Marcella Gomez

    Background: Wound transcriptomic analysis can be used to quantify wound healing stages and identify leverage points for wound healing intervention. However, individual gene signatures corresponding to wound healing stages vary from one experiment to another and are highly dependent on both experimental setup and bioinformatics methods. Methods: We develop a systematic approach to informatively compare time series from publicly available wound transcriptomic datasets, including mouse and human wounds, and identify consistent gene expression patterns. Results: We reveal the limitations of gene expression data collection, interpretation, and comparison. For example, the sample rate of wound transcriptomic sample collection must be higher than the rate of changes in the wound healing processes, otherwise, important changes in gene expression may be missed. This may lead to mis finding the most significant genes, as peaks of expression for highly differentially expressed genes are lost. Nevertheless, we derived a short list of genes highly differentially expressed in all datasets under consideration. After clustering and normalization, these genes clearly demonstrate similarly changing dynamics of expression between the wounds and may be used for wound healing stage detection.Conclusions: A list of genes that may be used for transcriptomics-based wound healing stage detection is provided. In addition, we suggest experimental approaches that could help researchers to extract more meaningful results.

  • Open Access

    Effective spatio-temporal regimes for wound treatment by way of macrophage polarization: A mathematical model

    by Jiahao Xue, Ksenia Zlobina, Marcella Gomez

    Wound healing consists of a sequence of biological processes often grouped into different stages. Interventions applied to accelerate normal wound healing must take into consideration timing with respect to wound healing stages in order to maximize treatment effectiveness. Macrophage polarization from M1 to M2 represents a transition from the inflammatory to proliferation stage of wound healing. Accelerating this transition may be an effective way to accelerate wound healing; however, it must be induced at the appropriate time. We search for an optimal spatio-temporal regime to apply wound healing treatment in a mathematical model of wound healing. In this work we show that to maximize effectiveness, treatment must not be applied too early or too late. We also show that effective spatial distribution of treatment depends on the heterogenity of the wound surface. In conclusion, this research provides a possible optimal regime of therapy that focuses on macrophage activity and a hypothesis of treatment outcome to be tested experimentally in future. Finding best regimes for treatment application is a first step towards development of intelligent algorithms of wound treatment that minimize healing time.