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

General General

Improved fully convolutional neuron networks on small retinal vessel segmentation using local phase as attention.

In Frontiers in medicine

Retinal images have been proven significant in diagnosing multiple diseases such as diabetes, glaucoma, and hypertension. Retinal vessel segmentation is crucial for the quantitative analysis of retinal images. However, current methods mainly concentrate on the segmentation performance of overall retinal vessel structures. The small vessels do not receive enough attention due to their small percentage in the full retinal images. Small retinal vessels are much more sensitive to the blood circulation system and have great significance in the early diagnosis and warning of various diseases. This paper combined two unsupervised methods, local phase congruency (LPC) and orientation scores (OS), with a deep learning network based on the U-Net as attention. And we proposed the U-Net using local phase congruency and orientation scores (UN-LPCOS), which showed a remarkable ability to identify and segment small retinal vessels. A new metric called sensitivity on a small ship (Sesv ) was also proposed to evaluate the methods' performance on the small vessel segmentation. Our strategy was validated on both the DRIVE dataset and the data from Maastricht Study and achieved outstanding segmentation performance on both the overall vessel structure and small vessels.

Kuang Xihe, Xu Xiayu, Fang Leyuan, Kozegar Ehsan, Chen Huachao, Sun Yue, Huang Fan, Tan Tao

2023

local phase, orientation scores, retinal vessel, segmentation, unsupervised enhancement

General General

Energy-based analog neural network framework.

In Frontiers in computational neuroscience

Over the past decade a body of work has emerged and shown the disruptive potential of neuromorphic systems across a broad range of studies, often combining novel machine learning models and nanotechnologies. Still, the scope of investigations often remains limited to simple problems since the process of building, training, and evaluating mixed-signal neural models is slow and laborious. In this paper, we introduce an open-source framework, called EBANA, that provides a unified, modularized, and extensible infrastructure, similar to conventional machine learning pipelines, for building and validating analog neural networks (ANNs). It uses Python as interface language with a syntax similar to Keras, while hiding the complexity of the underlying analog simulations. It already includes the most common building blocks and maintains sufficient modularity and extensibility to easily incorporate new concepts, electrical, and technological models. These features make EBANA suitable for researchers and practitioners to experiment with different design topologies and explore the various tradeoffs that exist in the design space. We illustrate the framework capabilities by elaborating on the increasingly popular Energy-Based Models (EBMs), used in conjunction with the local Equilibrium Propagation (EP) training algorithm. Our experiments cover 3 datasets having up to 60,000 entries and explore network topologies generating circuits in excess of 1,000 electrical nodes that can be extensively benchmarked with ease and in reasonable time thanks to the native EBANA parallelization capability.

Watfa Mohamed, Garcia-Ortiz Alberto, Sassatelli Gilles

2023

SPICE, analog, energy-based models, equilibrium propagation, framework, mixed-signal, neural networks

General General

Identifying the challenges and opportunities of PCOS awareness month by analysing its global digital impact.

In Frontiers in endocrinology ; h5-index 55.0

BACKGROUND AND OBJECTIVE : Although significant resources are invested each September for PCOS Awareness Month campaign, there are no studies measuring its impact. We evaluated the digital impact of PCOS Awareness Month, common themes and associated topics, top influencers, and global equity of influence during the PCOS Awareness month.

METHODS : In this serial cross-sectional analysis, we studied the outputs from Symplur® to study the total impressions of #PCOS on Twitter®. We tracked the hashtags-#PCOS, #PCOSawarenessmonth, #PCOSawareness-and a search query- "#PCOS OR #PCOSawarenessmonth OR #PCOSawareness"-using Sproutsocial® to study the total number of tweets related to PCOS Awareness Month. Network analysis was done using SocioViz® to identify common themes and associated topics. Using SymplurRank® machine learning algorithm, the top 10 #PCOS influencers were identified based on the number of mentions received. Google® Trends was used to study the web and news search popularity over the last 10 years beyond social media platforms.

RESULTS : An overall upward trend in the digital impact of PCOS awareness was noted since 2017. While the top themes associated with PCOS (insulin resistance, depression, anxiety, menopause, hormones, infertility) remained the same in 2021 and 2022, newer themes emerged in the latter year suggesting the need for ongoing review. News outlets were the most influential organisations during PCOS Awareness Month in both years of study. Seven of the top 10 users were the same in both years. Limited engagement from African, Asian, South American, and non-English speaking European countries was seen on Google Trends analysis.

CONCLUSION : Active involvement from various stakeholders of PCOS Awareness Month has shaped it into an effective strategy to raise awareness with social media playing a crucial role in amplifying the message. Our findings also provide an opportunity to understand the current perceptions and expectations amongst the public, which can influence future healthcare investment and research.

Malhotra Kashish, Pan Carina Synn Cuen, Davitadze Meri, Kempegowda Punith

2023

PCOS, PCOS awareness month, collaboration, global equity, mental health, polycystic ovary syndrome, social media

General General

Microwave-assisted valorization and characterization of Citrus limetta peel waste into pectin as a perspective food additive.

In Journal of food science and technology

Machine learning techniques were employed to evaluate the effect of process parameters viz. microwave power (100 W, 300 W, 600 W); pH (1, 1.5, 2); and microwave time (the 60 s, 120 s, 180 s) on the pectin yield from Citrus limetta peel. A fourth-order polynomial function of 66.60 scales was used by the Support Vector Regression (SVR) model at an epsilon (ε) value of 0.003. The co-efficient of determination (R2) and root mean square error-values for training data and test data were 0.984; 0.77 and 0.993; 0.66 respectively. At optimized conditions, microwave power 600 W, pH 1, and time 180 s the best yield of 32.75% was obtained. The integrity of pectin skeletal was confirmed with FTIR and 1H NMR spectrums. The physicochemical analysis revealed that CLP is a high-methoxyl pectin (HMP) with a 63.20 ± 0.88% degree of esterification, 798.45 ± 26.15 equivalent weight, 8.06 ± 0.62% methoxyl content, 67.93 ± 3.36 AUA content, 6.27 ± 0.27 g water/g pectin WHC, 2.68 ± 0.20 g oil/g pectin OHC, low moisture, ash and protein content of 6.85 ± 0.10%, 3.87 ± 0.10% and 2.61 ± 0.06% respectively, which can be utilized as a food additive. Therefore, pectin extraction from Citrus limetta peel using a greener technique like MAE is an eco-friendly, time-saving approach to transform waste into a versatile food additive.

Sharma Poonam, Osama Khwaja, Varjani Sunita, Farooqui Alvina, Younis Kaiser

2023-Apr

Eco-friendly, Food additive, Genetic algorithm, Microwave-assisted extraction, Pectin, Support vector regression

General General

Blood RNA alternative splicing events as diagnostic biomarkers for infectious disease.

In Cell reports methods

Assays detecting blood transcriptome changes are studied for infectious disease diagnosis. Blood-based RNA alternative splicing (AS) events, which have not been well characterized in pathogen infection, have potential normalization and assay platform stability advantages over gene expression for diagnosis. Here, we present a computational framework for developing AS diagnostic biomarkers. Leveraging a large prospective cohort of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and whole-blood RNA sequencing (RNA-seq) data, we identify a major functional AS program switch upon viral infection. Using an independent cohort, we demonstrate the improved accuracy of AS biomarkers for SARS-CoV-2 diagnosis compared with six reported transcriptome signatures. We then optimize a subset of AS-based biomarkers to develop microfluidic PCR diagnostic assays. This assay achieves nearly perfect test accuracy (61/62 = 98.4%) using a naive principal component classifier, significantly more accurate than a gene expression PCR assay in the same cohort. Therefore, our RNA splicing computational framework enables a promising avenue for host-response diagnosis of infection.

Zhang Zijun, Sauerwald Natalie, Cappuccio Antonio, Ramos Irene, Nair Venugopalan D, Nudelman German, Zaslavsky Elena, Ge Yongchao, Gaitas Angelo, Ren Hui, Brockman Joel, Geis Jennifer, Ramalingam Naveen, King David, McClain Micah T, Woods Christopher W, Henao Ricardo, Burke Thomas W, Tsalik Ephraim L, Goforth Carl W, Lizewski Rhonda A, Lizewski Stephen E, Weir Dawn L, Letizia Andrew G, Sealfon Stuart C, Troyanskaya Olga G

2023-Feb-27

RNA splicing, SARS-CoV-2, diagnostic biomarker, host response assays, infectious disease, viral infection

oncology Oncology

Machine learning approaches to predict drug efficacy and toxicity in oncology.

In Cell reports methods

In recent years, there has been a surge of interest in using machine learning algorithms (MLAs) in oncology, particularly for biomedical applications such as drug discovery, drug repurposing, diagnostics, clinical trial design, and pharmaceutical production. MLAs have the potential to provide valuable insights and predictions in these areas by representing both the disease state and the therapeutic agents used to treat it. To fully utilize the capabilities of MLAs in oncology, it is important to understand the fundamental concepts underlying these algorithms and how they can be applied to assess the efficacy and toxicity of therapeutics. In this perspective, we lay out approaches to represent both the disease state and the therapeutic agents used by MLAs to derive novel insights and make relevant predictions.

Badwan Bara A, Liaropoulos Gerry, Kyrodimos Efthymios, Skaltsas Dimitrios, Tsirigos Aristotelis, Gorgoulis Vassilis G

2023-Feb-27

artificial intelligence, drug discovery, drug response, machine learning