Kuzmenko, D. | Search bioRxiv

bioRxiv2023

Efficient Visual Perception of Human-Robot Walking Environments Using Semi-Supervised Learning

Convolutional neural networks trained using supervised learning can improve visual perception for human-robot walking. These advances have been possible due to large-scale datasets like ExoNet and StairNet - the largest open-source image datasets of real-world walking environments. However, these datasets require vast amounts of manually annotated data, the development of which is time consuming and labor intensive. Here we present a novel semi-supervised learning system (ExoNet-SSL) that uses over 1.2 million unlabelled images from ExoNet to improve training efficiency. We developed a deep learning model based on mobile vision transformers and trained the model using semi-supervised learning for image classification. Compared to standard supervised learning (98.4%), our ExoNet-SSL system was able to maintain high prediction accuracy (98.8%) when tested on previously unseen environments, while requiring 35% fewer labelled images during training. These results show that semi-supervised learning can improve training efficiency by leveraging large amounts of unlabelled data and minimize the size requirements for manually annotated images. Future research will focus on model deployment for onboard real-time inference and control of human-robot walking.

bioengineering↗

bioRxiv2020

Lipid binding by N-terminal motif mediates plasma membrane localization of Bordetella effector protein BteA

The classical Bordetella species, B. pertussis and B. bronchiseptica, employ a type III secretion system (T3SS) to inject a 69-kDa BteA effector into the host cells. Upon injection, BteA localizes to the cytosolic leaflet of lipid rafts via its N-terminal lipid raft targeting (LRT) domain and induces cell death. The plasma membrane targeting and cytotoxicity mechanisms of BteA are poorly understood. Using protein-lipid overlay assay and surface plasmon resonance, we showed here that the recombinant LRT domain, which adopts a four-helix bundle topology of membrane localization domains, specifically binds negatively charged membrane phospholipids. The binding affinity for phosphatidylinositol 4,5-bisphosphate (PIP2)-containing liposomes with Kd ~450 nM was higher than for those enriched in phosphatidylserine (Kd ~1.2 M) while both phospholipids were required for plasma membrane targeting in yeast cells. The membrane association of LRT further depended on its electrostatic and hydrophobic interactions and involved a loop L1-located leucine residue. Importantly, charge-reversal substitutions within the L1 region disrupted plasma membrane localization of BteA effector without hampering its cytotoxic activity during B. bronchiseptica infection of HeLa cells. The LRT-mediated targeting of BteA to the cytosolic leaflet of the plasma membrane of host cells is, hence, dispensable for the effector cytotoxicity. Author summaryThe respiratory pathogens of humans and other animals, Bordetella pertussis and Bordetella bronchiseptica, produce a type III secretion system effector protein BteA. This effector consists of two functional domains, an N-terminal lipid raft targeting (LRT) domain, and a cytotoxic C-terminal domain, which induces non-apoptotic and caspase-1-independent host cell death. We found here that the LRT domain of BteA associates with plasma membrane by binding to negatively charged phospholipids. We further discovered that the mechanism of membrane association by LRT is reminiscent of the one used by three other diverse families of toxins: clostridial glucosyltransferase toxins, multifunctional-autoprocessing RTX toxins (MARTX), and Pasteurella multocida-like toxins. Intriguingly, we also report that plasma membrane targeting by the LRT domain does not contribute to cytotoxic activity of BteA during B. bronchiseptica infection. Overall, our work elucidated the mechanism of plasma membrane association by LRT, and further provided the basis for future research on cellular activities of BteA and the mechanism of BteA-induced cell death.

microbiology↗