Tatikonda, R. R. | Search bioRxiv

bioRxiv2023

TCR-H: Machine Learning Prediction of T-cell Receptor Epitope Binding on Unseen Datasets

AI/ML approaches to predicting T-cell receptor (TCR) epitope specificity achieve high performance metrics on test datasets which include sequences that are also part of the training set but fail to generalize to test sets consisting of epitopes and TCRs that are absent from the training set, i.e., unseen. We present TCR-H, a supervised classification Support Vector Machines model using physicochemical features trained on the largest dataset available to date using only experimentally validated non-binders as negative datapoints. TCR-H exhibits an area under the curve of the receiver-operator characteristic (AUC of ROC) of 0.87 for epitope hard splitting (i.e., on test sets with all epitopes unseen), 0.92 for TCR hard splitting and 0.89 for strict splitting in which neither the epitopes nor the TCRs in the test set are seen in the training data. TCR-H may thus represent a significant step towards general applicability of epitope:TCR specificity prediction.

immunology↗

bioRxiv2020

Mechanical Activation of MscL Revealed by a Locally Distributed Tension Molecular Dynamics Approach

Membrane tension perceived by mechanosensitive (MS) proteins mediates cellular responses to mechanical stimuli and osmotic stresses, and it also guides multiple biological functions including cardiovascular control and development. In bacteria, MS channels function as tension-activated pores limiting excessive turgor pressure, with MscL (MS channel of large conductance) acting as an emergency release valve preventing cell lysis. Previous attempts to simulate gating transitions in MscL by either directly applying steering forces to the protein or by increasing the whole system tension were not fully successful and often disrupted the integrity of the system. We present a novel locally distributed tension molecular dynamics (LDT-MD) simulation method that allows application of forces continuously distributed among lipids surrounding the channel using a specially constructed collective variable. We report reproducible and reversible transitions of MscL to the open state with measured parameters of lateral expansion and conductivity that exactly satisfy experimental values. The LDT-MD method enables exploration of the MscL gating process with different pulling velocities and variable tension asymmetry between the inner and outer membrane leaflets. We use LDT-MD in combination with well-tempered metadynamics to reconstruct the tension-dependent free energy landscape for the opening transition in MscL. SIGNIFICANCEMembrane-embedded mechanosensitive (MS) proteins are essential for numerous biological functions including cardiovascular control and development, osmotic regulation, touch and pain sensing. In this work, we present a novel molecular dynamics simulation method that allows rapid and systematic exploration of structure, dynamics, and energetics of the mechanical transduction process in MS proteins under tightly controlled local tension distributed in the lipid rim around the protein. We provide a detailed description of the gating transition for the tension-activated bacterial mechanosensitive channel of large conductance, MscL, which is the best characterized channel of this type. MscL functions as a tension-activated emergency osmolyte release valve that limits excessive turgor pressure, prevents cell lysis and thus imparts environmental stability to most free-living bacteria.

biophysics↗