Schueffler, P. J. | bioRxiv Science

bioRxiv2024

MultiOmicsAgent: Guided extreme gradient-boosted decision trees-based approaches for biomarker-candidate discovery in multi-omics data

MultiOmicsAgent (MOAgent) is an innovative, Python based open-source tool for biomarker discovery, utilizing machine learning techniques specifically extreme gradient-boosted decision trees to process multi-omics data. With its cross-platform compatibility, user-oriented graphical interface and a well-documented API, MOAgent not only meets the needs of both coding professionals and those new to machine learning but also addresses common data analysis challenges like data incompleteness, class imbalances and data leakage between disjoint data splits. MOAgents guided data analysis strategy opens up data-driven insights from digitized clinical biospecimen cohorts and makes advanced data analysis accessible and reliable for a wide audience. Biographical NoteJens Settelmeier, Julia Boshart, Martin Gesell are Ph.D. candidates, Jianbo Fu, Sebastian N. Steiner are Post Doc candidates and Sandra Goetze, Patrick Pedrioli senior scientists at the Institute of Translational Medicine at Health Sciences and Technology department at ETH Zurich, Switzerland, within Professor Bernd Wollscheids research group who has been working in the fields of bioinformatics, clinical multi-omics with a focus on spatial cell surface proteomics. Peter J. Schuffler is professor at the institute of Pathology at the TU Munich, Germany and has been working in the field of digital pathology and clinical multi-modal studies. Diyora Salimova is junior professor at the department of Applied Mathematics at the Albert-Ludwigs-University of Freibug, Germany and has been working in the field of stochastic processes, approximation theory and machine learning related topics. Key PointsO_LIMOAgent enables a guided biomarker-candidate discovery in multi-omics studies, providing a graphical interface and well-documented API. C_LIO_LIA user can run MOAgent on a personal computer without the requirement of coding a single line. C_LIO_LIMOAgent is a Python-based solution for biomarker-candidate discovery, using machine learning to analyze multi-omics data. C_LIO_LIMOAgent can address challenges like data incompleteness and class imbalances, ensuring reliable analysis. C_LIO_LIMOAgent makes advanced data analysis accessible, enhancing insights from clinical data. C_LI

bioinformatics↗

bioRxiv2016

DeepScope: Nonintrusive Whole Slide Saliency Annotation and Prediction from Pathologists at the Microscope

Modern digital pathology departments have grown to produce whole-slide image data at petabyte scale, an unprecedented treasure chest for medical machine learning tasks. Unfortunately, most digital slides are not annotated at the image level, hindering large-scale application of supervised learning. Manual labeling is prohibitive, requiring pathologists with decades of training and outstanding clinical service responsibilities. This problem is further aggravated by the United States Food and Drug Administrations ruling that primary diagnosis must come from a glass slide rather than a digital image. We present the first end-to-end framework to overcome this problem, gathering annotations in a nonintrusive manner during a pathologists routine clinical work: (i) microscope-specific 3D-printed commodity camera mounts are used to video record the glass-slide-based clinical diagnosis process; (ii) after routine scanning of the whole slide, the video frames are registered to the digital slide; (iii) motion and observation time are estimated to generate a spatial and temporal saliency map of the whole slide. Demonstrating the utility of these annotations, we train a convolutional neural network that detects diagnosis-relevant salient regions, then report accuracy of 85.15% in bladder and 91.40% in prostate, with 75.00% accuracy when training on prostate but predicting in bladder, despite different pathologists examining the different tissues. When training on one patient but testing on another, AUROC in bladder is 0.7929 0.1109 and in prostate is 0.9568{+/-}0.0374. Our tool is available at https://bitbucket.org/aschaumberg/deepscope.

pathology↗