• Deep learning for early detection of pathological changes in X-ray bone microstructures: case of osteoarthritis

      Jakaite, Livija; Schetinin, Vitaly; Hladůvka, Jiří; Minaev, Sergey; Ambia, Aziz; Krzanowski, Wojtek; ; University of Bedfordshire; TU Wien; Stavropol State Medical University; et al. (Nature, 2021-01-27)
      Texture features are designed to quantitatively evaluate patterns of spatial distribution of image pixels for purposes of image analysis and interpretation. Unexplained variations in the texture patterns often lead to misinterpretation and undesirable consequences in medical image analysis. In this paper we explore the ability of machine learning (ML) methods to design a radiology test of Osteoarthritis (OA) at early stage when the number of patients’ cases is small. In our experiments we use high-resolution X-ray images of knees in patients which were identified with Kellgren–Lawrence scores progressing from 1. The existing ML methods have provided a limited diagnostic accuracy, whilst the proposed Group Method of Data Handling strategy of Deep Learning has significantly extended the diagnostic test. The comparative experiments demonstrate that the proposed framework using the Zernike-based texture features has significantly improved the diagnostic accuracy on average by 11%. This allows us to conclude that the designed model for early diagnostic of OA will provide more accurate radiology tests, although new study is required when a large number of patients’ cases will be available.
    • The intrinsically disordered Tarp protein from chlamydia binds actin with a partially preformed helix

      Tolchard, James; Walpole, Samuel J.; Miles, Andrew J.; Maytum, Robin; Eaglen, Lawrence A.; Hackstadt, Ted; Wallace, B.A.; Blumenschein, Tharin M.A. (Nature, 2018-01-31)
      Tarp (translocated actin recruiting phosphoprotein) is an effector protein common to all chlamydial species that functions to remodel the host-actin cytoskeleton during the initial stage of infection. In C. trachomatis, direct binding to actin monomers has been broadly mapped to a 100-residue region (726-825) which is predicted to be predominantly disordered, with the exception of a ~10-residue α-helical patch homologous to other WH2 actin-binding motifs. Biophysical investigations demonstrate that a Tarp726-825 construct behaves as a typical intrinsically disordered protein; within it, NMR relaxation measurements and chemical shift analysis identify the ten residue WH2-homologous region to exhibit partial α-helix formation. Isothermal titration calorimetry experiments on the same construct in the presence of monomeric G-actin show a well defined binding event with a 1:1 stoichiometry and Kd of 102 nM, whilst synchrotron radiation circular dichroism spectroscopy suggests the binding is concomitant with an increase in helical secondary structure. Furthermore, NMR experiments in the presence of G-actin indicate this interaction affects the proposed WH2-like α-helical region, supporting results from in silico docking calculations which suggest that, when folded, this α-helix binds within the actin hydrophobic cleft as seen for other actin-associated proteins.