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

Radiology Radiology

Artificial intelligence for assistance of radiology residents in chest CT evaluation for COVID-19 pneumonia: a comparative diagnostic accuracy study.

In Acta radiologica (Stockholm, Sweden : 1987)

BACKGROUND : In hospitals, it is crucial to rule out coronavirus disease 2019 (COVID-19) timely and reliably. Artificial intelligence (AI) provides sufficient accuracy to identify chest computed tomography (CT) scans with signs of COVID-19.

PURPOSE : To compare the diagnostic accuracy of radiologists with different levels of experience with and without assistance of AI in CT evaluation for COVID-19 pneumonia and to develop an optimized diagnostic pathway.

MATERIAL AND METHODS : The retrospective, single-center, comparative case-control study included 160 consecutive participants who had undergone chest CT scan between March 2020 and May 2021 without or with confirmed diagnosis of COVID-19 pneumonia in a ratio of 1:3. Index tests were chest CT evaluation by five radiological senior residents, five junior residents, and an AI software. Based on the diagnostic accuracy in every group and on comparison of groups, a sequential CT assessment pathway was developed.

RESULTS : Areas under receiver operating curves were 0.95 (95% confidence interval [CI]=0.88-0.99), 0.96 (95% CI=0.92-1.0), 0.77 (95% CI=0.68-0.86), and 0.95 (95% CI=0.9-1.0) for junior residents, senior residents, AI, and sequential CT assessment, respectively. Proportions of false negatives were 9%, 3%, 17%, and 2%, respectively. With the developed diagnostic pathway, junior residents evaluated all CT scans with the support of AI. Senior residents were only required as second readers in 26% (41/160) of the CT scans.

CONCLUSION : AI can support junior residents with chest CT evaluation for COVID-19 and reduce the workload of senior residents. A review of selected CT scans by senior residents is mandatory.

Mlynska Lucja, Malouhi Amer, Ingwersen Maja, Güttler Felix, Gräger Stephanie, Teichgräber Ulf

2023-Mar-08

Artificial intelligence, COVID-19, SARS-CoV-2, computed tomography, deep learning, neural networks

General General

Neuron Structure Modeling for Generalizable Remote Physiological Measurement

ArXiv Preprint

Remote photoplethysmography (rPPG) technology has drawn increasing attention in recent years. It can extract Blood Volume Pulse (BVP) from facial videos, making many applications like health monitoring and emotional analysis more accessible. However, as the BVP signal is easily affected by environmental changes, existing methods struggle to generalize well for unseen domains. In this paper, we systematically address the domain shift problem in the rPPG measurement task. We show that most domain generalization methods do not work well in this problem, as domain labels are ambiguous in complicated environmental changes. In light of this, we propose a domain-label-free approach called NEuron STructure modeling (NEST). NEST improves the generalization capacity by maximizing the coverage of feature space during training, which reduces the chance for under-optimized feature activation during inference. Besides, NEST can also enrich and enhance domain invariant features across multi-domain. We create and benchmark a large-scale domain generalization protocol for the rPPG measurement task. Extensive experiments show that our approach outperforms the state-of-the-art methods on both cross-dataset and intra-dataset settings.

Hao Lu, Zitong Yu, Xuesong Niu, Yingcong Chen

2023-03-10

General General

Evaluating network-based missing protein prediction using p-values, Bayes Factors, and probabilities.

In Journal of bioinformatics and computational biology

Some prediction methods use probability to rank their predictions, while some other prediction methods do not rank their predictions and instead use [Formula: see text]-values to support their predictions. This disparity renders direct cross-comparison of these two kinds of methods difficult. In particular, approaches such as the Bayes Factor upper Bound (BFB) for [Formula: see text]-value conversion may not make correct assumptions for this kind of cross-comparisons. Here, using a well-established case study on renal cancer proteomics and in the context of missing protein prediction, we demonstrate how to compare these two kinds of prediction methods using two different strategies. The first strategy is based on false discovery rate (FDR) estimation, which does not make the same naïve assumptions as BFB conversions. The second strategy is a powerful approach which we colloquially call "home ground testing". Both strategies perform better than BFB conversions. Thus, we recommend comparing prediction methods by standardization to a common performance benchmark such as a global FDR. And where this is not possible, we recommend reciprocal "home ground testing".

Goh Wilson Wen Bin, Kong Weijia, Wong Limsoon

2023-Mar-09

Bayes factors, PROTREC, Statistics, [Formula: see text]-values, data science, machine learning, missing proteins, networks

Radiology Radiology

Predicting depressed and elevated mood symptomatology in bipolar disorder using brain functional connectomes.

In Psychological medicine ; h5-index 82.0

BACKGROUND : The study is aimed to identify brain functional connectomes predictive of depressed and elevated mood symptomatology in individuals with bipolar disorder (BD) using the machine learning approach Connectome-based Predictive Modeling (CPM).

METHODS : Functional magnetic resonance imaging data were obtained from 81 adults with BD while they performed an emotion processing task. CPM with 5000 permutations of leave-one-out cross-validation was applied to identify functional connectomes predictive of depressed and elevated mood symptom scores on the Hamilton Depression and Young Mania rating scales. The predictive ability of the identified connectomes was tested in an independent sample of 43 adults with BD.

RESULTS : CPM predicted the severity of depressed [concordance between actual and predicted values (r = 0.23, pperm (permutation test) = 0.031) and elevated (r = 0.27, pperm = 0.01) mood. Functional connectivity of left dorsolateral prefrontal cortex and supplementary motor area nodes, with inter- and intra-hemispheric connections to other anterior and posterior cortical, limbic, motor, and cerebellar regions, predicted depressed mood severity. Connectivity of left fusiform and right visual association area nodes with inter- and intra-hemispheric connections to the motor, insular, limbic, and posterior cortices predicted elevated mood severity. These networks were predictive of mood symptomatology in the independent sample (r ⩾ 0.45, p = 0.002).

CONCLUSIONS : This study identified distributed functional connectomes predictive of depressed and elevated mood severity in BD. Connectomes subserving emotional, cognitive, and psychomotor control predicted depressed mood severity, while those subserving emotional and social perceptual functions predicted elevated mood severity. Identification of these connectome networks may help inform the development of targeted treatments for mood symptoms.

Sankar Anjali, Shen Xilin, Colic Lejla, Goldman Danielle A, Villa Luca M, Kim Jihoon A, Pittman Brian, Scheinost Dustin, Constable R Todd, Blumberg Hilary P

2023-Mar-09

Bipolar disorder, CPM, functional magnetic resonance imaging, symptom severity

General General

Higher-order modular regulation of the human proteome.

In Molecular systems biology

Operons are transcriptional modules that allow bacteria to adapt to environmental changes by coordinately expressing the relevant set of genes. In humans, biological pathways and their regulation are more complex. If and how human cells coordinate the expression of entire biological processes is unclear. Here, we capture 31 higher-order co-regulation modules, which we term progulons, by help of supervised machine-learning on proteomics data. Progulons consist of dozens to hundreds of proteins that together mediate core cellular functions. They are not restricted to physical interactions or co-localisation. Progulon abundance changes are primarily controlled at the level of protein synthesis and degradation. Implemented as a web app at www.proteomehd.net/progulonFinder, our approach enables the targeted search for progulons of specific cellular processes. We use it to identify a DNA replication progulon and reveal multiple new replication factors, validated by extensive phenotyping of siRNA-induced knockdowns. Progulons provide a new entry point into the molecular understanding of biological processes.

Kustatscher Georg, Hödl Martina, Rullmann Edward, Grabowski Piotr, Fiagbedzi Emmanuel, Groth Anja, Rappsilber Juri

2023-Mar-09

DNA replication, mRNA coexpression, machine-learning, protein co-regulation, quantitative proteomics

General General

Unsupervised Data-Driven Reconstruction of Molecular Motifs in Simple to Complex Dynamic Micelles.

In The journal of physical chemistry. B

The reshuffling mobility of molecular building blocks in self-assembled micelles is a key determinant of many their interesting properties, from emerging morphologies and surface compartmentalization, to dynamic reconfigurability and stimuli-responsiveness. However, the microscopic details of such complex structural dynamics are typically nontrivial to elucidate, especially in multicomponent assemblies. Here we show a machine-learning approach that allows us to reconstruct the structural and dynamic complexity of mono- and bicomponent surfactant micelles from high-dimensional data extracted from equilibrium molecular dynamics simulations. Unsupervised clustering of smooth overlap of atomic position (SOAP) data enables us to identify, in a set of multicomponent surfactant micelles, the dominant local molecular environments that emerge within them and to retrace their dynamics, in terms of exchange probabilities and transition pathways of the constituent building blocks. Tested on a variety of micelles differing in size and in the chemical nature of the constitutive self-assembling units, this approach effectively recognizes the molecular motifs populating them in an exquisitely agnostic and unsupervised way, and allows correlating them to their composition in terms of constitutive surfactant species.

Cardellini Annalisa, Crippa Martina, Lionello Chiara, Afrose Syed Pavel, Das Dibyendu, Pavan Giovanni M

2023-Mar-08