Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

General General

In vivo identification of apoptotic and extracellular vesicle-bound live cells using image-based deep learning.

In Journal of extracellular vesicles

The in vivo detection of dead cells remains a major challenge due to technical hurdles. Here, we present a novel method, where injection of fluorescent milk fat globule-EGF factor 8 protein (MFG-E8) in vivo combined with imaging flow cytometry and deep learning allows the identification of dead cells based on their surface exposure of phosphatidylserine (PS) and other image parameters. A convolutional autoencoder (CAE) was trained on defined pictures and successfully used to identify apoptotic cells in vivo. However, unexpectedly, these analyses also revealed that the great majority of PS+ cells were not apoptotic, but rather live cells associated with PS+ extracellular vesicles (EVs). During acute viral infection apoptotic cells increased slightly, while up to 30% of lymphocytes were decorated with PS+ EVs of antigen-presenting cell (APC) exosomal origin. The combination of recombinant fluorescent MFG-E8 and the CAE-method will greatly facilitate analyses of cell death and EVs in vivo.

Kranich Jan, Chlis Nikolaos-Kosmas, Rausch Lisa, Latha Ashretha, Schifferer Martina, Kurz Tilman, Foltyn-Arfa Kia Agnieszka, Simons Mikael, Theis Fabian J, Brocker Thomas

2020-Jul-16

Extracellular Vesicles, apoptosis, dendritic cells, exosomes, irradiation, viral Infection

Surgery Surgery

Machine intelligence for nerve conduit design and production.

In Journal of biological engineering

Nerve guidance conduits (NGCs) have emerged from recent advances within tissue engineering as a promising alternative to autografts for peripheral nerve repair. NGCs are tubular structures with engineered biomaterials, which guide axonal regeneration from the injured proximal nerve to the distal stump. NGC design can synergistically combine multiple properties to enhance proliferation of stem and neuronal cells, improve nerve migration, attenuate inflammation and reduce scar tissue formation. The aim of most laboratories fabricating NGCs is the development of an automated process that incorporates patient-specific features and complex tissue blueprints (e.g. neurovascular conduit) that serve as the basis for more complicated muscular and skin grafts. One of the major limitations for tissue engineering is lack of guidance for generating tissue blueprints and the absence of streamlined manufacturing processes. With the rapid expansion of machine intelligence, high dimensional image analysis, and computational scaffold design, optimized tissue templates for 3D bioprinting (3DBP) are feasible. In this review, we examine the translational challenges to peripheral nerve regeneration and where machine intelligence can innovate bottlenecks in neural tissue engineering.

Stewart Caleb E, Kan Chin Fung Kelvin, Stewart Brody R, Sanicola Henry W, Jung Jangwook P, Sulaiman Olawale A R, Wang Dadong

2020

Artificial intelligence, Bioprinting, Computer vision, Data science, Machine learning, Nerve regeneration, Tissue engineering

Ophthalmology Ophthalmology

Laplacian feature detection and feature alignment for multimodal ophthalmic image registration using phase correlation and Hessian affine feature space.

In Signal processing

Advances in multimodal imaging have revolutionized diagnostic and treatment monitoring in ophthalmic practice. In multimodal ophthalmic imaging, geometric deformations are inevitable and they contain inherent deformations arising from heterogeneity in the optical characteristics of imaging devices and patient related factors. The registration of ophthalmic images under such conditions is challenging. We propose a novel technique that overcomes these challenges, using Laplacian feature, Hessian affine feature space and phase correlation, to register blue autofluorescence, near-infrared reflectance and color fundus photographs of the ocular posterior pole with high accuracy. Our validation analysis - that used current feature detection and extraction techniques (speed-up robust features (SURF), a concept of wind approach (KAZE), and fast retina keypoint (FREAK)), and quantitative measures (Sørensen-Dice coefficient, Jaccard index, and Kullback-Leibler divergence scores) - showed that our approach has significant merit in registering multimodal images when compared with a mix-and-match SURF-KAZE-FREAK benchmark approach. Similarly, our evaluation analysis that used a state-of-the-art qualitative measure - the mean registration error (MRE) - showed that the proposed approach is significantly better than the mix-and-match SURF-KAZE-FREAK benchmark approach, as well as a cutting edge image registration technique - Linear Stack Alignment with SIFT (scale-invariant feature transform) - in registering multimodal ophthalmic images.

Suthaharan Shan, Rossi Ethan A, Snyder Valerie, Chhablani Jay, Lejoyeux Raphael, Sahel Jośe-Alain, Dansingani Kunal

2020-Dec

2010 MSC: 68U10, 92C55, 94A12, 97R40, Hessian feature space, computational models, image registration, machine learning, multimodal imaging, ophthalmology, phase correlation

Public Health Public Health

Evaluating the informativeness of deep learning annotations for human complex diseases.

In Nature communications ; h5-index 260.0

Deep learning models have shown great promise in predicting regulatory effects from DNA sequence, but their informativeness for human complex diseases is not fully understood. Here, we evaluate genome-wide SNP annotations from two previous deep learning models, DeepSEA and Basenji, by applying stratified LD score regression to 41 diseases and traits (average N = 320K), conditioning on a broad set of coding, conserved and regulatory annotations. We aggregated annotations across all (respectively blood or brain) tissues/cell-types in meta-analyses across all (respectively 11 blood or 8 brain) traits. The annotations were highly enriched for disease heritability, but produced only limited conditionally significant results: non-tissue-specific and brain-specific Basenji-H3K4me3 for all traits and brain traits respectively. We conclude that deep learning models have yet to achieve their full potential to provide considerable unique information for complex disease, and that their conditional informativeness for disease cannot be inferred from their accuracy in predicting regulatory annotations.

Dey Kushal K, van de Geijn Bryce, Kim Samuel Sungil, Hormozdiari Farhad, Kelley David R, Price Alkes L

2020-Sep-17

General General

Reinforcing materials modelling by encoding the structures of defects in crystalline solids into distortion scores.

In Nature communications ; h5-index 260.0

This work revises the concept of defects in crystalline solids and proposes a universal strategy for their characterization at the atomic scale using outlier detection based on statistical distances. The proposed strategy provides a generic measure that describes the distortion score of local atomic environments. This score facilitates automatic defect localization and enables a stratified description of defects, which allows to distinguish the zones with different levels of distortion within the structure. This work proposes applications for advanced materials modelling ranging from the surrogate concept for the energy per atom to the relevant information selection for evaluation of energy barriers from the mean force. Moreover, this concept can serve for design of robust interatomic machine learning potentials and high-throughput analysis of their databases. The proposed definition of defects opens up many perspectives for materials design and characterization, promoting thereby the development of novel techniques in materials science.

Goryaeva Alexandra M, Lapointe Clovis, Dai Chendi, Dérès Julien, Maillet Jean-Bernard, Marinica Mihai-Cosmin

2020-Sep-17

General General

Large-scale prediction and analysis of protein sub-mitochondrial localization with DeepMito.

In BMC bioinformatics

BACKGROUND : The prediction of protein subcellular localization is a key step of the big effort towards protein functional annotation. Many computational methods exist to identify high-level protein subcellular compartments such as nucleus, cytoplasm or organelles. However, many organelles, like mitochondria, have their own internal compartmentalization. Knowing the precise location of a protein inside mitochondria is crucial for its accurate functional characterization. We recently developed DeepMito, a new method based on a 1-Dimensional Convolutional Neural Network (1D-CNN) architecture outperforming other similar approaches available in literature.

RESULTS : Here, we explore the adoption of DeepMito for the large-scale annotation of four sub-mitochondrial localizations on mitochondrial proteomes of five different species, including human, mouse, fly, yeast and Arabidopsis thaliana. A significant fraction of the proteins from these organisms lacked experimental information about sub-mitochondrial localization. We adopted DeepMito to fill the gap, providing complete characterization of protein localization at sub-mitochondrial level for each protein of the five proteomes. Moreover, we identified novel mitochondrial proteins fishing on the set of proteins lacking any subcellular localization annotation using available state-of-the-art subcellular localization predictors. We finally performed additional functional characterization of proteins predicted by DeepMito as localized into the four different sub-mitochondrial compartments using both available experimental and predicted GO terms. All data generated in this study were collected into a database called DeepMitoDB (available at http://busca.biocomp.unibo.it/deepmitodb ), providing complete functional characterization of 4307 mitochondrial proteins from the five species.

CONCLUSIONS : DeepMitoDB offers a comprehensive view of mitochondrial proteins, including experimental and predicted fine-grain sub-cellular localization and annotated and predicted functional annotations. The database complements other similar resources providing characterization of new proteins. Furthermore, it is also unique in including localization information at the sub-mitochondrial level. For this reason, we believe that DeepMitoDB can be a valuable resource for mitochondrial research.

Savojardo Castrense, Martelli Pier Luigi, Tartari Giacomo, Casadio Rita

2020-Sep-16

Convolutional neural network, Deep learning, Functional annotation, Mitochondrial protein, Subcellular localization, Submitochondrial localization