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

General General

A review of modern technologies for tackling COVID-19 pandemic.

In Diabetes & metabolic syndrome

OBJECTIVE : Science and technology sector constituting of data science, machine learning and artificial intelligence are contributing towards COVID-19. The aim of the present study is to discuss the various aspects of modern technology used to fight against COVID-19 crisis at different scales, including medical image processing, disease tracking, prediction outcomes, computational biology and medicines.

METHODS : A progressive search of the database related to modern technology towards COVID-19 is made. Further, a brief review is done on the extracted information by assessing the various aspects of modern technologies for tackling COVID-19 pandemic.

RESULTS : We provide a window of thoughts on review of the technology advances used to decrease and smother the substantial impact of the outburst. Though different studies relating to modern technology towards COVID-19 have come up, yet there are still constrained applications and contributions of technology in this fight.

CONCLUSIONS : On-going progress in the modern technology has contributed in improving people's lives and hence there is a solid conviction that validated research plans including artificial intelligence will be of significant advantage in helping people to fight this infection.

Kumar Aishwarya, Gupta Puneet Kumar, Srivastava Ankita


Artificial intelligence, COVID-19, Epidemic, Machine learning

General General

Local features and global shape information in object classification by deep convolutional neural networks.

In Vision research ; h5-index 38.0

Deep convolutional neural networks (DCNNs) show impressive similarities to the human visual system. Recent research, however, suggests that DCNNs have limitations in recognizing objects by their shape. We tested the hypothesis that DCNNs are sensitive to an object's local contour features but have no access to global shape information that predominates human object recognition. We employed transfer learning to assess local and global shape processing in trained networks. In Experiment 1, we used restricted and unrestricted transfer learning to retrain AlexNet, VGG-19, and ResNet-50 to classify circles and squares. We then probed these networks with stimuli with conflicting global shape and local contour information. We presented networks with overall square shapes comprised of curved elements and circles comprised of corner elements. Networks classified the test stimuli by local contour features rather than global shapes. In Experiment 2, we changed the training data to include circles and squares comprised of different elements so that the local contour features of the object were uninformative. This considerably increased the network's tendency to produce global shape responses, but deeper analyses in Experiment 3 revealed the network still showed no sensitivity to the spatial configuration of local elements. These findings demonstrate that DCNNs' performance is an inversion of human performance with respect to global and local shape processing. Whereas abstract relations of elements predominate in human perception of shape, DCNNs appear to extract only local contour fragments, with no representation of how they spatially relate to each other to form global shapes.

Baker Nicholas, Lu Hongjing, Erlikhman Gennady, Kellman Philip J


Deep learning, Global and local features, Object recognition, Shape

General General

General Health Questionnaire (GHQ-12), Beck Depression Inventory (BDI-6), and Mental Health Index (MHI-5): psychometric and predictive properties in a Finnish population-based sample.

In Psychiatry research ; h5-index 64.0

The short versions of the General Health Questionnaire (GHQ-12), Beck's Depression Inventory (BDI-6), and Mental Health Index (MHI-5) are all valid and reliable measures of general psychological distress, depressive symptoms, and anxiety. We tested the psychometric properties of the scales, their overlap, and their ability to predict mental health service use using both regression and machine learning (ML, random forest) approaches. Data were from the population-based FinHealth-2017 Study of adults (N = 4270) with data on all of the evaluated instruments. Constructive validity, internal consistency, invariance, and optimal cut-off points in predicting mental health services were tested. Constructive validity was acceptable and all instruments measured their own distinct phenomenon. Some of the item scoring in BDI-6 was not optimal, and the sensitivity and specificity of all scales were relatively weak in predicting service use. Small gender differences emerged in optimal cut-off points. ML did not improve model predictions. GHQ-12, BDI-6, and MHI-5 may be interpreted to measure different constructs of psychological health symptoms, but are not particularly useful predictors of service use.

Elovanio Marko, Hakulinen Christian, Pulkki-Råback Laura, Aalto Anna-Mari, Virtanen Marianna, Partonen Timo, Suvisaari Jaana


Depression, Distress, Measuring, Population, Psychometrics

oncology Oncology

Epithelial-to-mesenchymal transition is a prognostic marker for patient outcome in advanced stage HNSCC patients treated with chemoradiotherapy.

In Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology

BACKGROUND : The prognosis of patients with HPV-negative advanced stage head and neck squamous cell carcinoma (HNSCC) remains poor. No prognostic markers other than TNM staging are routinely used in clinic. Epithelial-to-mesenchymal transition (EMT) has been shown to be a strong prognostic factor in other cancer types. The purpose of this study was to determine the role of EMT in HPV-negative HNSCC outcomes.

METHODS : Pretreatment tumor material from patients of two cohorts, totaling 174 cisplatin-based chemoradiotherapy treated HPV-negative HNSCC patients, was RNA-sequenced. Seven different EMT gene expression signatures were used for EMT status classification and generation of HNSCC-specific EMT models using Random Forest machine learning.

RESULTS : Mesenchymal classification by all EMT signatures consistently enriched for poor prognosis patients in both cohorts of 98 and 76 patients. Uni- and multivariate analyses show important HR of 1.6-5.8, thereby revealing EMT's role in HNSCC outcome. Discordant classification by these signatures prompted the generation of an HNSCC-specific EMT profile based on the concordantly classified samples in the first cohort (cross-validation AUC>0.98). The independent validation cohort confirmed the association of mesenchymal classification by the HNSCC-EMT model with poor overall survival (HR=3.39, p<0.005) and progression free survival (HR=3.01, p<0.005) in multivariate analysis with TNM. Analysis of an additional HNSCC cohort from PET-positive patients with metastatic disease prior to treatment further supports this relationship and reveals a strong link of EMT to the propensity to metastasize.

CONCLUSIONS : EMT in HPV-negative HNSCC co-defines patient outcome after chemoradiotherapy. The generated HNSCC-EMT prediction models can function as strong prognostic biomarkers.

van der Heijden Martijn, Essers Paul B M, Verhagen Caroline V M, Willems Stefan M, Sanders Joyce, de Roest Reinout H, Vossen David M, Leemans C René, Verheij Marcel, Brakenhoff Ruud H, van den Brekel Michiel W M, Vens Conchita


Chemoradiotherapy, Epithelial to mesenchymal transition, HNSCC, Head and neck cancer, Prognostic biomarkers, RNA-Seq

General General

Procleave: Predicting Protease-specific Substrate Cleavage Sites by Combining Sequence and Structural Information.

In Genomics, proteomics & bioinformatics

Proteases are enzymes that cleave and hydrolyse the peptide bonds between two specific amino acid residues of target substrate proteins. Protease-controlled proteolysis plays a key role in the degradation and recycling of proteins, which is essential for various physiological processes. Thus, solving the substrate identification problem will have important implications for the precise understanding of functions and physiological roles of proteases, as well as for therapeutic target identification and pharmaceutical applicability. Consequently, there is a great demand for bioinformatics methods that can predict novel substrate cleavage events with high accuracy by utilizing both sequence and structural information. In this study, we present Procleave, a novel bioinformatics approach for predicting protease-specific substrates and specific cleavage sites by taking into account both their sequence and 3D structural information. Structural features of known cleavage sites were represented by discrete values using a LOWESS data-smoothing optimization method, which turned out to be critical for the performance of Procleave. The optimal approximations of all structural parameter values were encoded in a conditional random field (CRF) computational framework, alongside sequence and chemical group-based features. Here, we demonstrate the outstanding performance of Procleave through extensive benchmarking and independent tests. Procleave is capable of correctly identifying most cleavage sites in the case study. Importantly, when applied to the human structural proteome encompassing 17,628 protein structures, Procleave suggests a number of potential novel target substrates and their corresponding cleavage sites of different proteases. Procleave is implemented as a webserver and is freely accessible at

Li Fuyi, Leier Andre, Liu Quanzhong, Wang Yanan, Xiang Dongxu, Akutsu Tatsuya, Webb Geoffrey I, Ian Smith A, Marquez-Lago Tatiana, Li Jian, Song Jiangning


Cleavage site prediction, Conditional random field, Machine learning, Protease, Structural determinants

Surgery Surgery

Automated Laparoscopic Colorectal Surgery Workflow Recognition using Artificial Intelligence: Experimental Research.

In International journal of surgery (London, England)

BACKGROUND : Identifying laparoscopic surgical videos using artificial intelligence (AI) facilitates the automation of several currently time-consuming manual processes, including video analysis, indexing, and video-based skill assessment. This study aimed to construct a large annotated dataset comprising laparoscopic colorectal surgery (LCRS) videos from multiple institutions and evaluate the accuracy of automatic recognition for surgical phase, action, and tool by combining this dataset with AI.

MATERIALS AND METHODS : A total of 300 intraoperative videos were collected from 19 high-volume centers. A series of surgical workflows were classified into 9 phases and 3 actions, and the area of 5 tools were assigned by painting. More than 82 million frames were annotated for a phase and action classification task, and 4,000 frames were annotated for a tool segmentation task. Of these frames, 80% were used for the training dataset and 20% for the test dataset. A convolutional neural network (CNN) was used to analyze the videos. Intersection over union (IoU) was used as the evaluation metric for tool recognition.

RESULTS : The overall accuracies for the automatic surgical phase and action classification task were 81.0% and 83.2%, respectively. The mean IoU for the automatic tool segmentation task for 5 tools was 51.2%.

CONCLUSIONS : A large annotated dataset of LCRS videos was constructed, and the phase, action, and tool were recognized with high accuracy using AI. Our dataset has potential uses in medical applications such as automatic video indexing and surgical skill assessments. Open research will assist in improving CNN models by making our dataset available in the field of computer vision.

Kitaguchi Daichi, Takeshita Nobuyoshi, Matsuzaki Hiroki, Oda Tatsuya, Watanabe Masahiko, Mori Kensaku, Kobayashi Etsuko, Ito Masaaki


artificial intelligence, automatic video indexing, convolutional neural network, laparoscopic colorectal surgery, surgical skill assessment, surgical workflow recognition