Associations of Maternal Breastmilk microRNAs and Infant Obesity Status at 1 Year.

Infant consumption of human milk (HM) is associated with a reduced risk of overweight and obesity, but the reasons for this relationship are not completely understood. There is emerging evidence that micro RNAs (miRNAs) regulate infant development and metabolism, but the associations between HM miRNAs and infant growth remain poorly understood. We examined the relationship between HM miRNA consumption and infant obesity in 163 mother-infant dyads to determine (1) if miRNA profiles differentiate infants with obesity, and (2) if individual miRNAs accurately predicted infant obesity status at one year of age. Infant obesity was categorized as weight-for-length (WFL) Z scores or conditional weight gain (CWG) in the 95th percentile. HM miRNA profile was associated with infant age (r2 = 6.4%, p = 0.001), but not maternal obesity status (r2 = 1.5%, p = 0.87) or infant weight status (WFL Z-score) at birth (r2 = 0.6%, p = 0.4), 1 month (r2 = 0.5%, p = 0.6), or 4 months (r2 = 0.8%, p = 0.2). Nine HM miRNAs were associated with either 12-month CWG or 12-month WFL Z scores. Among these 9 miRNAs, miR-224-5p remained significant in a logistic regression model that accounted for additional demographic factors (estimate = -27.57, p = 0.004). These findings suggest involvement of HM miRNAs and particularly miR-224-5p in infant growth, warranting further investigation. To our knowledge, this is the largest study of HM miRNAs and early-life obesity and contributes to the understanding of the relationship between HM miRNAs and infant growth.

Milk levels of transforming growth factor beta 1 identify mothers with low milk supply.

Human milk is optimal for infant nutrition. However, many mothers cease breastfeeding because of low milk supply (LMS). It is difficult to identify mothers at risk for LMS because its biologic underpinnings are not fully understood. Previously, we demonstrated that milk micro-ribonucleic acids (miRNAs) may be related to LMS. Transforming growth factor beta (TGFß) also plays an important role in mammary involution and may contribute to LMS. We performed a longitudinal cohort study of 139 breastfeeding mothers to test the hypothesis that milk levels of TGFß would identify mothers with LMS. We explored whether TGFß impacts the expression of LMS-related miRNAs in cultured human mammary epithelial cells (HMECs). LMS was defined by maternal report of inadequate milk production, and confirmed by age of formula introduction and infant weight trajectory. Levels of TGF-ß1 and TGF-ß2 were measured one month after delivery. There was a significant relationship between levels of TGF-ß1 and LMS (X2 = 8.92, p = 0.003) on logistic regression analysis, while controlling for lactation stage (X2 = 1.28, p = 0.25), maternal pre-pregnancy body mass index (X2 = 0.038, p = 0.84), and previous breastfeeding experience (X2 = 7.43, p = 0.006). The model accounted for 16.8% of variance in the data (p = 0.005) and correctly predicted LMS for 84.6% of mothers (22/26; AUC = 0.72). Interactions between TGF-ß1 and miR-22-3p displayed significant effect on LMS status (Z = 2.67, p = 0.008). Further, incubation of HMECs with TGF-ß1 significantly reduced mammary cell number (t = -4.23, p = 0.003) and increased levels of miR-22-3p (t = 3.861, p = 0.008). Interactions between TGF-ß1 and miR-22-3p may impact mammary function and milk levels of TGF-ß1 could have clinical utility for identifying mothers with LMS. Such information could be used to provide early, targeted lactation support.

Using an AI-based avatar for interviewer training at Children's Advocacy Centers: Proof of Concept.

This proof-of- concept study focused on interviewers' behaviors and perceptions when interacting with a dynamic AI child avatar alleging abuse. Professionals (N = 68) took part in a virtual reality (VR) study in which they questioned an avatar presented as a child victim of sexual or physical abuse. Of interest was how interviewers questioned the avatar, how productive the child avatar was in response, and how interviewers perceived the VR interaction. Findings suggested alignment between interviewers' virtual questioning approaches and interviewers' typical questioning behavior in real-world investigative interviews, with a diverse range of questions used to elicit disclosures from the child avatar. The avatar responded to most question types as children typically do, though more nuanced programming of the avatar's productivity in response to complex question types is needed. Participants rated the avatar positively and felt comfortable with the VR experience. Results underscored the potential of AI-based interview training as a scalable, standardized alternative to traditional methods.

A large-scale multivariate soccer athlete health, performance, and position monitoring dataset.

Data analysis for athletic performance optimization and injury prevention is of tremendous interest to sports teams and the scientific community. However, sports data are often sparse and hard to obtain due to legal restrictions, unwillingness to share, and lack of personnel resources to be assigned to the tedious process of data curation. These constraints make it difficult to develop automated systems for analysis, which require large datasets for learning. We therefore present SoccerMon, the largest soccer athlete dataset available today containing both subjective and objective metrics, collected from two different elite women's soccer teams over two years. Our dataset contains 33,849 subjective reports and 10,075 objective reports, the latter including over six billion GPS position measurements. SoccerMon can not only play a valuable role in developing better analysis and prediction systems for soccer, but also inspire similar data collection activities in other domains which can benefit from subjective athlete reports, GPS position information, and/or time-series data in general.

Visual explanations for polyp detection: How medical doctors assess intrinsic versus extrinsic explanations.

Deep learning has achieved immense success in computer vision and has the potential to help physicians analyze visual content for disease and other abnormalities. However, the current state of deep learning is very much a black box, making medical professionals skeptical about integrating these methods into clinical practice. Several methods have been proposed to shed some light on these black boxes, but there is no consensus on the opinion of medical doctors that will consume these explanations. This paper presents a study asking medical professionals about their opinion of current state-of-the-art explainable artificial intelligence methods when applied to a gastrointestinal disease detection use case. We compare two different categories of explanation methods, intrinsic and extrinsic, and gauge their opinion of the current value of these explanations. The results indicate that intrinsic explanations are preferred and that physicians see value in the explanations. Based on the feedback collected in our study, future explanations of medical deep neural networks can be tailored to the needs and expectations of doctors. Hopefully, this will contribute to solving the issue of black box medical systems and lead to successful implementation of this powerful technology in the clinic.

A dataset for predicting cloud cover over Europe.

Clouds are important factors when projecting future climate. Unfortunately, future cloud fractional cover (the portion of the sky covered by clouds) is associated with significant uncertainty, making climate projections difficult. In this paper, we present the European Cloud Cover dataset, which can be used to learn statistical relations between cloud cover and other environmental variables, to potentially improve future climate projections. The dataset was created using a novel technique called Area Weighting Regridding Scheme to map satellite observations to cloud fractional cover on the same grid as the other variables in the dataset. Baseline experiments using autoregressive models document that it is possible to use the dataset to predict cloud fractional cover.

Assessing generalisability of deep learning-based polyp detection and segmentation methods through a computer vision challenge.

Polyps are well-known cancer precursors identified by colonoscopy. However, variability in their size, appearance, and location makes the detection of polyps challenging. Moreover, colonoscopy surveillance and removal of polyps are highly operator-dependent procedures and occur in a highly complex organ topology. There exists a high missed detection rate and incomplete removal of colonic polyps. To assist in clinical procedures and reduce missed rates, automated methods for detecting and segmenting polyps using machine learning have been achieved in past years. However, the major drawback in most of these methods is their ability to generalise to out-of-sample unseen datasets from different centres, populations, modalities, and acquisition systems. To test this hypothesis rigorously, we, together with expert gastroenterologists, curated a multi-centre and multi-population dataset acquired from six different colonoscopy systems and challenged the computational expert teams to develop robust automated detection and segmentation methods in a crowd-sourcing Endoscopic computer vision challenge. This work put forward rigorous generalisability tests and assesses the usability of devised deep learning methods in dynamic and actual clinical colonoscopy procedures. We analyse the results of four top performing teams for the detection task and five top performing teams for the segmentation task. Our analyses demonstrate that the top-ranking teams concentrated mainly on accuracy over the real-time performance required for clinical applicability. We further dissect the devised methods and provide an experiment-based hypothesis that reveals the need for improved generalisability to tackle diversity present in multi-centre datasets and routine clinical procedures.

CELLULAR, A Cell Autophagy Imaging Dataset.

Cells in living organisms are dynamic compartments that continuously respond to changes in their environment to maintain physiological homeostasis. While basal autophagy exists in cells to aid in the regular turnover of intracellular material, autophagy is also a critical cellular response to stress, such as nutritional depletion. Conversely, the deregulation of autophagy is linked to several diseases, such as cancer, and hence, autophagy constitutes a potential therapeutic target. Image analysis to follow autophagy in cells, especially on high-content screens, has proven to be a bottleneck. Machine learning (ML) algorithms have recently emerged as crucial in analyzing images to efficiently extract information, thus contributing to a better understanding of the questions at hand. This paper presents CELLULAR, an open dataset consisting of images of cells expressing the autophagy reporter mRFP-EGFP-Atg8a with cell-specific segmentation masks. Each cell is annotated into either basal autophagy, activated autophagy, or unknown. Furthermore, we introduce some preliminary experiments using the dataset that can be used as a baseline for future research.

Sperm motility assessed by deep convolutional neural networks into WHO categories.

Semen analysis is central in infertility investigation. Manual assessment of sperm motility according to the WHO recommendations is the golden standard, and extensive training is a requirement for accurate and reproducible results. Deep convolutional neural networks (DCNN) are especially suitable for image classification. In this study, we evaluated the performance of the DCNN ResNet-50 in predicting the proportion of sperm in the WHO motility categories. Two models were evaluated using tenfold cross-validation with 65 video recordings of wet semen preparations from an external quality assessment programme for semen analysis. The corresponding manually assessed data was obtained from several of the reference laboratories, and the mean values were used for training of the DCNN models. One model was trained to predict the three categories progressive motility, non-progressive motility, and immotile spermatozoa. Another model was used in predicting four categories, where progressive motility was differentiated into rapid and slow. The resulting average mean absolute error (MAE) was 0.05 and 0.07, and the average ZeroR baseline was 0.09 and 0.10 for the three-category and the four-category model, respectively. Manual and DCNN-predicted motility was compared by Pearson's correlation coefficient and by difference plots. The strongest correlation between the mean manually assessed values and DCNN-predicted motility was observed for % progressively motile spermatozoa (Pearson's r = 0.88, p < 0.001) and % immotile spermatozoa (r = 0.89, p < 0.001). For rapid progressive motility, the correlation was moderate (Pearson's r = 0.673, p < 0.001). The median difference between manual and predicted progressive motility was 0 and 2 for immotile spermatozoa. The largest bias was observed at high and low percentages of progressive and immotile spermatozoa. The DCNN-predicted value was within the range of the interlaboratory variation of the results for most of the samples. In conclusion, DCNN models were able to predict the proportion of spermatozoa into the WHO motility categories with significantly lower error than the baseline. The best correlation between the manual and the DCNN-predicted motility values was found for the categories progressive and immotile. Of note, there was considerable variation between the mean motility values obtained for each category by the reference laboratories, especially for rapid progressive motility, which impacts the training of the DCNN models.

Infant Saliva Microbiome Activity Modulates Nutritional Impacts on Neurodevelopment.

Neurodevelopment is influenced by complex interactions between environmental factors, including social determinants of health (SDOH), nutrition, and even the microbiome. This longitudinal cohort study of 142 infants tested the hypothesis that microbial activity modulates the effects of nutrition on neurodevelopment. Salivary microbiome activity was measured at 6 months using RNA sequencing. Infant nutrition was assessed longitudinally with the Infant Feeding Practices survey. The primary outcome was presence/absence of neurodevelopmental delay (NDD) at 18 months on the Survey of Wellbeing in Young Children. A logistic regression model employing two microbial factors, one nutritional factor, and two SDOH accounted for 33.3% of the variance between neurodevelopmental groups (p < 0.001, AIC = 77.7). NDD was associated with Hispanic ethnicity (OR 18.1, 2.36-139.3; p = 0.003), no fish consumption (OR 10.6, 2.0-54.1; p = 0.003), and increased Candidatus Gracilibacteria activity (OR 1.43, 1.00-2.07; p = 0.007). Home built after 1977 (OR 0.02, 0.001-0.53; p = 0.004) and Chlorobi activity (OR 0.76, 0.62-0.93, p = 0.001) were associated with reduced risk of NDD. Microbial alpha diversity modulated the effect of fish consumption on NDD (X2 = 5.7, p = 0.017). These data suggest the benefits of fish consumption for neurodevelopment may be mediated by microbial diversity. Confirmation in a larger, randomized trial is required.

Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis.

Deep neural networks are complex machine learning models that have shown promising results in analyzing high-dimensional data such as those collected from medical examinations. Such models have the potential to provide fast and accurate medical diagnoses. However, the high complexity makes deep neural networks and their predictions difficult to understand. Providing model explanations can be a way of increasing the understanding of "black box" models and building trust. In this work, we applied transfer learning to develop a deep neural network to predict sex from electrocardiograms. Using the visual explanation method Grad-CAM, heat maps were generated from the model in order to understand how it makes predictions. To evaluate the usefulness of the heat maps and determine if the heat maps identified electrocardiogram features that could be recognized to discriminate sex, medical doctors provided feedback. Based on the feedback, we concluded that, in our setting, this mode of explainable artificial intelligence does not provide meaningful information to medical doctors and is not useful in the clinic. Our results indicate that improved explanation techniques that are tailored to medical data should be developed before deep neural networks can be applied in the clinic for diagnostic purposes.

Increased Sleep Latency and Decreased Sleep Duration are Associated With Elevated Risk of Bed Sharing Among Mother-Infant Dyads.

Bed sharing increases risk of sleep-related infant deaths. We hypothesized that infant sleep difficulties increase bed sharing, independent of social determinants of health (SDOH). In total, 191 mother-infant dyads in a prospective, longitudinal cohort study completed the Brief Infant Sleep Questionnaire at 1, 4, 6, and 12 months. Sleep characteristics at 1 month (latency, duration, night awakenings) were compared between dyads with/without bed sharing in the first 12 months. Infants who participated in bed sharing slept fewer hours at night (7.1 ± 1.7 hours vs 8.3 ± 1.5 hours, P = .001, d = -0.79), and took longer to fall asleep (0.7 ± 0.6 hours vs 0.5 ± 0.5 hours, P = .021, d = 0.43), even when controlling for SDOH variables that influence bed sharing. Maternal perception of sleep problems did not differ between groups (P = .12). Our findings suggest that infants with quantifiable sleep difficulties at 1 month are more likely to bed share.

Saliva microRNA Profile in Children with and without Severe SARS-CoV-2 Infection.

Severe acute respiratory syndrome corona virus 2 (SARS-CoV-2) may impair immune modulating host microRNAs, causing severe disease. Our objectives were to determine the salivary miRNA profile in children with SARS-CoV-2 infection at presentation and compare the expression in those with and without severe outcomes. Children <18 years with SARS-CoV-2 infection evaluated at two hospitals between March 2021 and February 2022 were prospectively enrolled. Severe outcomes included respiratory failure, shock or death. Saliva microRNAs were quantified with RNA sequencing. Data on 197 infected children (severe = 45) were analyzed. Of the known human miRNAs, 1606 (60%) were measured and compared across saliva samples. There were 43 miRNAs with ≥2-fold difference between severe and non-severe cases (adjusted p-value < 0.05). The majority (31/43) were downregulated in severe cases. The largest between-group differences involved miR-4495, miR-296-5p, miR-548ao-3p and miR-1273c. These microRNAs displayed enrichment for 32 gene ontology pathways including viral processing and transforming growth factor beta and Fc-gamma receptor signaling. In conclusion, salivary miRNA levels are perturbed in children with severe COVID-19, with the majority of miRNAs being down regulated. Further studies are required to validate and determine the utility of salivary miRNAs as biomarkers of severe COVID-19.

VISEM-Tracking, a human spermatozoa tracking dataset.

A manual assessment of sperm motility requires microscopy observation, which is challenging due to the fast-moving spermatozoa in the field of view. To obtain correct results, manual evaluation requires extensive training. Therefore, computer-aided sperm analysis (CASA) has become increasingly used in clinics. Despite this, more data is needed to train supervised machine learning approaches in order to improve accuracy and reliability in the assessment of sperm motility and kinematics. In this regard, we provide a dataset called VISEM-Tracking with 20 video recordings of 30 seconds (comprising 29,196 frames) of wet semen preparations with manually annotated bounding-box coordinates and a set of sperm characteristics analyzed by experts in the domain. In addition to the annotated data, we provide unlabeled video clips for easy-to-use access and analysis of the data via methods such as self- or unsupervised learning. As part of this paper, we present baseline sperm detection performances using the YOLOv5 deep learning (DL) model trained on the VISEM-Tracking dataset. As a result, we show that the dataset can be used to train complex DL models to analyze spermatozoa.

The Association between Infant Colic and the Multi-Omic Composition of Human Milk.

Infant colic is a common condition with unclear biologic underpinnings and limited treatment options. We hypothesized that complex molecular networks within human milk (i.e., microbes, micro-ribonucleic acids (miRNAs), cytokines) would contribute to colic risk, while controlling for medical, social, and nutritional variables. This hypothesis was tested in a cohort of 182 breastfed infants, assessed with a modified Infant Colic Scale at 1 month. RNA sequencing was used to interrogate microbial and miRNA features. Luminex assays were used to measure growth factors and cytokines. Milk from mothers of infants with colic (n = 28) displayed higher levels of Staphylococcus (adj. p = 0.038, d = 0.30), miR-224-3p (adj. p = 0.023, d = 0.33), miR-125b-5p (adj. p = 0.028, d = 0.29), let-7a-5p (adj. p = 0.028, d = 0.27), and miR-205-5p (adj. p = 0.029, d = 0.26) compared to milk from non-colic mother-infant dyads (n = 154). Colic symptom severity was directly associated with milk hepatocyte growth factor levels (R = 0.21, p = 0.025). A regression model involving let-7a-5p, miR-29a-3p, and Lactobacillus accurately modeled colic risk (X2 = 16.7, p = 0.001). Molecular factors within human milk may impact colic risk, and provide support for a dysbiotic/inflammatory model of colic pathophysiology.

Human Milk-Derived Levels of let-7g-5p May Serve as a Diagnostic and Prognostic Marker of Low Milk Supply in Breastfeeding Women.

Low milk supply (LMS) is associated with early breastfeeding cessation; however, the biological underpinnings in the mammary gland are not understood. MicroRNAs (miRNAs) are small non-coding RNAs that post-transcriptionally downregulate gene expression, and we hypothesized the profile of miRNAs secreted into milk reflects lactation performance. Longitudinal changes in milk miRNAs were measured using RNAseq in women with LMS (n = 47) and adequate milk supply (AMS; n = 123). Relationships between milk miRNAs, milk supply, breastfeeding outcomes, and infant weight gain were assessed, and interactions between milk miRNAs, maternal diet, smoking status, and BMI were determined. Women with LMS had lower milk volume (p = 0.003), were more likely to have ceased breast feeding by 24 wks (p = 0.0003) and had infants with a lower mean weight-for-length z-score (p = 0.013). Milk production was significantly associated with milk levels of miR-16-5p (R = -0.14, adj p = 0.044), miR-22-3p (R = 0.13, adj p = 0.044), and let-7g-5p (R = 0.12, adj p = 0.046). Early milk levels of let-7g-5p were significantly higher in mothers with LMS (adj p = 0.0025), displayed an interaction between lactation stage and milk supply (p < 0.001), and were negatively related to fruit intake (p = 0.015). Putative targets of let-7g-5p include genes important to hormone signaling, RNA regulation, ion transport, and the extracellular matrix, and down-regulation of two targets (PRLR and IGF2BP1/IMP1) was confirmed in mammary cells overexpressing let-7g-5p in vitro. Our data provide evidence that milk-derived miRNAs reflect lactation performance in women and warrant further investigation to assess their utility for predicting LMS risk and early breastfeeding cessation.

Comparison of Symptom Duration Between Children With SARS-CoV-2 and Peers With Other Viral Illnesses During the COVID-19 Pandemic.

Some children and young people (CYP) with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) experience persistent symptoms, commonly called "long COVID." It remains unclear whether symptoms of SARS-CoV-2 persist longer than those of other respiratory viruses, particularly in young children. This cross-sectional study involved 372 CYP (0-15 years) tested for SARS-CoV-2. Character and duration of symptoms (cough, runny nose, sore throat, rash, diarrhea, vomiting, sore muscles, fatigue, fever, loss of smell) were compared between CYP with a positive test (n = 100) and those with a negative test (n = 272), while controlling for medical/demographic covariates. The average duration of symptoms for CYP with a positive SARS-CoV-2 test (8.5 ± 10 days) did not differ from that of CYP with a negative test (7.2 ± 5 days, P = .71, d = 0.046). A positive SARS-CoV-2 test did not increase the risk (36/372, 10%) of symptoms persisting for ≥3 weeks (odds ratio = 0.96, 95% confidence interval = 0.45-2.0). These results suggest CYP with non-SARS-CoV-2 infections experience a similar duration of symptoms as peers with SARS-CoV-2 infection.

Infant Saliva Levels of microRNA miR-151a-3p Are Associated with Risk for Neurodevelopmental Delay.

Prompt recognition of neurodevelopmental delay is critical for optimizing developmental trajectories. Currently, this is achieved with caregiver questionnaires whose sensitivity and specificity can be limited by socioeconomic and cultural factors. This prospective study of 121 term infants tested the hypothesis that microRNA measurement could aid early recognition of infants at risk for neurodevelopmental delay. Levels of four salivary microRNAs implicated in childhood autism (miR-125a-5p, miR-148a-5p, miR-151a-3p, miR-28-3p) were measured at 6 months of age, and compared between infants who displayed risk for neurodevelopmental delay at 18 months (n = 20) and peers with typical development (n = 101), based on clinical evaluation aided by the Survey of Wellbeing in Young Children (SWYC). Accuracy of microRNAs for predicting neurodevelopmental concerns at 18 months was compared to the clinical standard (9-month SWYC). Infants with neurodevelopmental concerns at 18 months displayed higher levels of miR-125a-5p (d = 0.30, p = 0.018, adj p = 0.049), miR-151a-3p (d = 0.30, p = 0.017, adj p = 0.048), and miR-28-3p (d = 0.31, p = 0.014, adj p = 0.048). Levels of miR-151a-3p were associated with an 18-month SWYC score (R = -0.19, p = 0.021) and probability of neurodevelopmental delay at 18 months (OR = 1.91, 95% CI, 1.14-3.19). Salivary levels of miR-151a-3p enhanced predictive accuracy for future neurodevelopmental delay (p = 0.010, X2 = 6.71, AUC = 0.71) compared to the 9-month SWYC score alone (OR = 0.56, 95% CI, 0.20-1.58, AUC = 0.567). This pilot study provides evidence that miR-151a-3p may aid the identification of infants at risk for neurodevelopmental delay. External validation of these findings is necessary.

Multi-Omic Factors Associated with Frequency of Upper Respiratory Infections in Developing Infants.

Susceptibility to upper respiratory infections (URIs) may be influenced by host, microbial, and environmental factors. We hypothesized that multi-omic analyses of molecular factors in infant saliva would identify complex host-environment interactions associated with URI frequency. A cohort study involving 146 infants was used to assess URI frequency in the first year of life. Saliva was collected at 6 months for high-throughput multi-omic measurement of cytokines, microRNAs, transcripts, and microbial RNA. Regression analysis identified environmental (daycare attendance, atmospheric pollution, breastfeeding duration), microbial (Verrucomicrobia, Streptococcus phage), and host factors (miR-22-5p) associated with URI frequency (p < 0.05). These results provide pathophysiologic clues about molecular factors that influence URI susceptibility. Validation of these findings in a larger cohort could one day yield novel approaches to detecting and managing URI susceptibility in infants.