Molecular dissection of an immunodominant epitope in Kv1.2-exclusive autoimmunity.

Introduction: Subgroups of autoantibodies directed against voltage-gated potassium channel (Kv) complex components have been associated with immunotherapy-responsive clinical syndromes. The high prevalence and the role of autoantibodies directly binding Kv remain, however, controversial. Our objective was to determine Kv autoantibody binding requirements and to clarify their contribution to the observed immune response. Methods: Binding epitopes were studied in sera (n = 36) and cerebrospinal fluid (CSF) (n = 12) from a patient cohort positive for Kv1.2 but negative for 32 common neurological autoantigens and controls (sera n = 18 and CSF n = 5) by phospho and deep mutational scans. Autoantibody specificity and contribution to the observed immune response were resolved on recombinant cells, cerebellum slices, and nerve fibers. Results: 83% of the patients (30/36) within the studied cohort shared one out of the two major binding epitopes with Kv1.2-3 reactivity. Eleven percent (4/36) of the serum samples showed no binding. Fingerprinting resolved close to identical sequence requirements for both shared epitopes. Kv autoantibody response is directed against juxtaparanodal regions in peripheral nerves and the axon initial segment in central nervous system neurons and exclusively mediated by the shared epitopes. Discussion: Systematic mapping revealed two shared autoimmune responses, with one dominant Kv1.2-3 autoantibody epitope being unexpectedly prevalent. The conservation of the molecular binding requirements among these patients indicates a uniform autoantibody repertoire with monospecific reactivity. The enhanced sensitivity of the epitope-based (10/12) compared with that of the cell-based detection (7/12) highlights its use for detection. The determined immunodominant epitope is also the primary immune response visible in tissue, suggesting a diagnostic significance and a specific value for routine screening.

Data augmentation in natural language processing: a novel text generation approach for long and short text classifiers.

In many cases of machine learning, research suggests that the development of training data might have a higher relevance than the choice and modelling of classifiers themselves. Thus, data augmentation methods have been developed to improve classifiers by artificially created training data. In NLP, there is the challenge of establishing universal rules for text transformations which provide new linguistic patterns. In this paper, we present and evaluate a text generation method suitable to increase the performance of classifiers for long and short texts. We achieved promising improvements when evaluating short as well as long text tasks with the enhancement by our text generation method. Especially with regard to small data analytics, additive accuracy gains of up to 15.53% and 3.56% are achieved within a constructed low data regime, compared to the no augmentation baseline and another data augmentation technique. As the current track of these constructed regimes is not universally applicable, we also show major improvements in several real world low data tasks (up to +4.84 F1-score). Since we are evaluating the method from many perspectives (in total 11 datasets), we also observe situations where the method might not be suitable. We discuss implications and patterns for the successful application of our approach on different types of datasets.

Inline application of digital holography [Invited].

We describe the inline integration of the digital holographic sensor HoloTop in a precision turning plant. A fully automated part-handling system that fulfills the requirements for cycle time and stability was built and integrated into the production process. The inspection system has been running in multishift operation since 2015. For the first time, to the best of our knowledge, the results of one-year, long-term height measurements of 10 million parts under rough production conditions are presented to verify the suitability for industrial use.