Interaction with intestinal microbes in infancy has a profound impact on health and disease in later life through programming of immune and metabolic pathways. We collected maternal faeces, placenta, amniotic fluid, colostrum, meconium and infant faeces samples from 15 mother-infant pairs in an effort to rigorously investigate prenatal and neonatal microbial transfer and gut colonisation. To ensure sterile sampling, only deliveries at full term by elective caesarean section were studied. Microbiota composition and activity assessment by conventional bacterial culture, 16S rRNA gene pyrosequencing, quantitative PCR, and denaturing gradient gel electrophoresis revealed that the placenta and amniotic fluid harbour a distinct microbiota characterised by low richness, low diversity and the predominance of Proteobacteria. Shared features between the microbiota detected in the placenta and amniotic fluid and in infant meconium suggest microbial transfer at the foeto-maternal interface. At the age of 3–4 days, the infant gut microbiota composition begins to resemble that detected in colostrum. Based on these data, we propose that the stepwise microbial gut colonisation process may be initiated already prenatally by a distinct microbiota in the placenta and amniotic fluid. The link between the mother and the offspring is continued after birth by microbes present in breast milk.

Breast feeding results in long term health benefits in the prevention of communicable and non-communicable diseases at both individual and population levels. Geographical location directly impacts the composition of breast milk including microbiota and lipids. The aim of this study was to investigate the influence of geographical location, i.e., Europe (Spain and Finland), Africa (South Africa), and Asia (China), on breast milk microbiota and lipid composition in samples obtained from healthy mothers after the 1 month of lactation. Altogether, 80 women (20 from each country) participated in the study, with equal number of women who delivered by vaginal or cesarean section from each country. Lipid composition particularly that of polyunsaturated fatty acids differed between the countries, with the highest amount of n-6 PUFA (25.6%) observed in the milk of Chinese women. Milk microbiota composition also differed significantly between the countries (p = 0.002). Among vaginally delivered women, Spanish women had highest amount of Bacteroidetes (mean relative abundance of 3.75) whereas Chinese women had highest amount of Actinobacteria (mean relative abundance 5.7). Women who had had a cesarean section had higher amount of Proteobacteria as observed in the milk of the Spanish and South African women. Interestingly, the Spanish and South African women had significantly higher bacterial genes mapped to lipid, amino acid and carbohydrate metabolism (p < 0.05). Association of the lipid profile with the microbiota revealed that monounsaturated fatty acids (MUFA) were negatively associated with Proteobacteria (r = -0.43, p < 0.05), while Lactobacillus genus was associated with MUFA (r = -0.23, p = 0.04). These findings reveal that the milk microbiota and lipid composition exhibit differences based on geographical locations in addition to the differences observed due to the mode of delivery.

Human milk oligosaccharides (HMOs) are structurally diverse unconjugated glycans with a composition unique to each lactating mother. While HMOs have been shown to have an impact on the development of infant gut microbiota, it is not well known if HMOs also already affect milk microbial composition. To address this question, we analysed eleven colostrum samples for HMO content by high-pressure liquid chromatography and microbiota composition by quantitative PCR. Higher total HMO concentration was associated with higher counts of Bifidobacterium spp. (ρ=0.63, P=0.036). A distinctive effect was seen when comparing different HMO groups: positive correlations were observed between sialylated HMOs and Bifidobacterium breve (ρ=0.84, P=0.001), and non-fucosylated/non-sialylated HMOs and Bifidobacterium longum group (ρ=0.65, P=0.030). In addition to associations between HMOs and bifidobacteria, positive correlations were observed between fucosylated HMOs and Akkermansia muciniphila (ρ=0.70, P=0.017), and between fucosylated/sialylated HMOs and Staphylococcus aureus (ρ=0.75, P=0.007). Our results suggest that the characterised HMOs have an effect on specific microbial groups in human milk. Both oligosaccharides and microbes provide a concise inoculum for the compositional development of the infant gut microbiota.

The identification of sleep stages is essential in the diagnostics of sleep disorders, among which obstructive sleep apnea (OSA) is one of the most prevalent. However, manual scoring of sleep stages is time-consuming, subjective, and costly. To overcome this shortcoming, we aimed to develop an accurate deep learning approach for automatic classification of sleep stages and to study the effect of OSA severity on the classification accuracy. Overnight polysomnographic recordings from a public dataset of healthy individuals (Sleep-EDF, n = 153) and from a clinical dataset (n = 891) of patients with

Study Objectives Accurate identification of sleep stages is essential in the diagnosis of sleep disorders (e.g. obstructive sleep apnea [OSA]) but relies on labor-intensive electroencephalogram (EEG)-based manual scoring. Furthermore, long-term assessment of sleep relies on actigraphy differentiating only between wake and sleep periods without identifying specific sleep stages and having low reliability in identifying wake periods after sleep onset. To address these issues, we aimed to develop an automatic method for identifying the sleep stages from the photoplethysmogram (PPG) signal obtained with a simple finger pulse oximeter. Methods PPG signals from the diagnostic polysomnographies of susptected OSA patients (n = 894) were utilized to develop a combined convolutional and recurrent neural network. The deep learning model was trained individually for three-stage (wake/NREM/REM), four-stage (wake/N1+N2/N3/REM), and five-stage (wake/N1/N2/N3/REM) classification of sleep. Results The three-stage model achieved an epoch-by-epoch accuracy of 80.1% with Cohen’s κ of 0.65. The four- and five-stage models achieved 68.5% (κ = 0.54), and 64.1% (κ = 0.51) accuracies, respectively. With the five-stage model, the total sleep time was underestimated with a mean bias error (SD) of of 7.5 (55.2) minutes. Conclusion The PPG-based deep learning model enabled accurate estimation of sleep time and differentiation between sleep stages with a moderate agreement to manual EEG-based scoring. As PPG is already included in ambulatory polygraphic recordings, applying the PPG-based sleep staging could improve their diagnostic value by enabling simple, low-cost, and reliable monitoring of sleep and help assess otherwise overlooked conditions such as REM-related OSA.