Search

Background Current copy number variation (CNV) identification methods have rapidly become mature. However, the postdetection processes such as variant interpretation or reporting are inefficient. To overcome this situation, we developed REDBot as an automated software package for accurate and direct generation of clinical diagnostic reports for prenatal and products of conception (POC) samples. Methods We applied natural language process (NLP) methods for analyzing 30,235 in‐house historical clinical reports through active learning, and then, developed clinical knowledge bases, evidence‐based interpretation methods and reporting criteria to support the whole postdetection pipeline. Results Of the 30,235 reports, we obtained 37,175 CNV‐paragraph pairs. For these pairs, the active learning approaches achieved a 0.9466 average F1‐score in sentence classification. The overall accuracy for variant classification was 95.7%, 95.2%, and 100.0% in retrospective, prospective, and clinical utility experiments, respectively. Conclusion By integrating NLP methods in CNVs postdetection pipeline, REDBot is a robust and rapid tool with clinical utility for prenatal and POC diagnosis.

Person search aims at localizing and identifying a query person from a gallery of uncropped scene images. Different from person re-identification (re-ID), its performance also depends on the localization accuracy of a pedestrian detector. The state-of-the-art methods train the detector individually, and the detected bounding boxes may be sub-optimal for the following re-ID task. To alleviate this issue, we propose a re-ID driven localization refinement framework for providing the refined detection boxes for person search. Specifically, we develop a differentiable ROI transform layer to effectively transform the bounding boxes from the original images. Thus, the box coordinates can be supervised by the re-ID training other than the original detection task. With this supervision, the detector can generate more reliable bounding boxes, and the downstream re-ID model can produce more discriminative embeddings based on the refined person localizations. Extensive experimental results on the widely used benchmarks demonstrate that our proposed method performs favorably against the state-of-the-art person search methods., 10 pages, 7 figures. Accepted by ICCV 2019

The estimation of 3D human body pose and shape from a single image has been extensively studied in recent years. However, the texture generation problem has not been fully discussed. In this paper, we propose an end-to-end learning strategy to generate textures of human bodies under the supervision of person re-identification. We render the synthetic images with textures extracted from the inputs and maximize the similarity between the rendered and input images by using the re-identification network as the perceptual metrics. Experiment results on pedestrian images show that our model can generate the texture from a single image and demonstrate that our textures are of higher quality than those generated by other available methods. Furthermore, we extend the application scope to other categories and explore the possible utilization of our generated textures., Comment: Accepted by CVPR2019

SummaryImprovements in shotgun proteomics approaches are hampered by increases in multiplexed (chimeric) spectra, as improvements in peak capacity, sensitivity or dynamic range all increase the number of co-eluting peptides. This results in diminishing returns using traditional search algorithms, as co-fragmented spectra are known to decrease identification rates. Here we describe MSCypher, a freely available software suite that enables an extensible workflow including a hybrid supervised machine learned strategy that dynamically adjusts to individual datasets. This results in improved identification rates and quantification of low-abundant peptides and proteins. In addition, the integration of peptide de novo sequencing and database searching enables an unbiased view of variants and high-intensity unassigned peptide spectral matches.HighlightsOpen-source end-to-end label-free proteomics workflowIntegrated database searching and machine learningCustomisable and extensible workflow including de novo sequencingOptimised for multiplexed spectra, challenging proteomics datasets and peptidomics applications