Neighborhood Co-regularized Multi-view Spectral Clustering of Microbiome Data

Adaptively Weighted Multiview Proximity Learning for Clustering

Recently, the proximity-based methods have achieved great success for multiview clustering. Nevertheless, most existing proximity-based methods take the predefined proximity matrices as input and their performance relies heavily on the quality of the predefined proximity matrices. A few multiview proximity learning (MVPL) methods have been proposed to tackle this problem but there are still some limitations, such as only emphasizing the intraview relation but overlooking the inter-view correlation, or not taking the weight differences of different views into account when considering the inter-view correlation. These limitations affect the quality of the learned proximity matrices and therefore influence the clustering performance. With the aim of breaking through these limitations simultaneously, a novel proximity learning method, called adaptively weighted MVPL (AWMVPL), is proposed. In the proposed method, both the intraview relation and the inter-view correlation are considered. Besides, when considering the inter-view correlation, the weights of different views are learned in a self-weighted scheme. Furthermore, through an adaptively weighted scheme, the information of the learned view-specific proximity matrices is integrated into a view-common cluster indicator matrix which outputs the final clustering result. Extensive experiments are conducted on several synthetic and real-world datasets to demonstrate the effectiveness and superiority of our method compared with the existing methods.

Generative Partial Multi-View Clustering With Adaptive Fusion and Cycle Consistency

Nowadays, with the rapid development of data collection sources and feature extraction methods, multi-view data are getting easy to obtain and have received increasing research attention in recent years, among which, multi-view clustering (MVC) forms a mainstream research direction and is widely used in data analysis. However, existing MVC methods mainly assume that each sample appears in all the views, without considering the incomplete view case due to data corruption, sensor failure, equipment malfunction, etc. In this study, we design and build a generative partial multi-view clustering model with adaptive fusion and cycle consistency, named as GP-MVC, to solve the incomplete multi-view problem by explicitly generating the data of missing views. The main idea of GP-MVC lies in two-fold. First, multi-view encoder networks are trained to learn common low-dimensional representations, followed by a clustering layer to capture the shared cluster structure across multiple views. Second, view-specific generative adversarial networks with multi-view cycle consistency are developed to generate the missing data of one view conditioning on the shared representation given by other views. These two steps could be promoted mutually, where the learned common representation facilitates data imputation and the generated data could further explores the view consistency. Moreover, an weighted adaptive fusion scheme is implemented to exploit the complementary information among different views. Experimental results on four benchmark datasets are provided to show the effectiveness of the proposed GP-MVC over the state-of-the-art methods.

At least three phenotypes exist among periodontitis patients

AIM

To identify phenotypes of periodontitis patients by the use of an unsupervised modelling technique (clustering), based on pre-treatment radiographic and microbiological characteristics.

MATERIALS AND METHODS

This retrospective study included data from 392 untreated periodontitis patients. Co-regularized spectral clustering algorithm was used to cluster the patients. The resulting clusters were subsequently characterized based on their demographics, radiographic bone loss patterns and microbial data.

RESULTS

The vast majority of patients fitted into one of the three main clusters (accuracy 90%). Cluster A (n = 18) was characterized by high prevalence and high proportions of Aggregatibacter actinomycetemcomitans, a trend for a more localized pattern of alveolar bone loss and young individuals. Clusters B (n = 200) and C (n = 135) differed clearly in disease severity patterns and smoking habits, but not in microbiological characteristics.

CONCLUSION

On the basis of alveolar bone loss patterns and microbiological data, untreated periodontitis patients can be clustered into at least three phenotypes. These results should be validated in other cohorts, and the clinical utility needs to be explored on the basis of periodontal treatment outcomes and/or disease progression.

Reversal of visceral hypersensitivity in rat by Menthacarin® , a proprietary combination of essential oils from peppermint and caraway, coincides with mycobiome modulation

BACKGROUND

Irritable bowel syndrome (IBS) is a common gastrointestinal disorder associated with altered gastrointestinal microflora and increased nociception to colonic distension. This visceral hypersensitivity can be reversed in our rat maternal separation model by fungicides. Menthacarin^® is a proprietary combination of essential oils from Mentha x piperita L. and Carum carvi. Because these oils exhibit antifungal and antibacterial properties, we investigated whether Menthacarin^® can reverse existing visceral hypersensitivity in maternally separated rats.

METHODS

In non-handled and maternally separated rats, we used the visceromotor responses to colorectal distension as measure for visceral sensitivity. We evaluated this response before and 24 hours after water-avoidance stress and after 7 days treatment with Menthacarin^® or control. The pre- and post-treatment mycobiome and microbiome were characterized by sequencing of fungal internal transcribed spacer 1 (ITS-1) and bacterial 16s rDNA regions. In vitro antifungal and antimicrobial properties of Menthacarin^® were studied with radial diffusion assay.

KEY RESULTS

Menthacarin^® inhibited in vitro growth of yeast and bacteria. Water-avoidance caused visceral hypersensitivity in maternally separated rats, and this was reversed by treatment. Multivariate analyses of ITS-1 and 16S high throughput data showed that maternal separation, induced changes in the myco- and microbiome. Menthacarin^® treatment of non-handled and maternally separated rats shifted the mycobiomes to more similar compositions.

CONCLUSIONS & INFERENCES

The development of visceral hypersensitivity in maternally separated rats and the Menthacarin^® -mediated reversal of hypersensitivity is associated with changes in the mycobiome. Therefore, Menthacarin^® may be a safe and effective treatment option that should be tested for IBS.

A Study of the Variation in the Salivary Peptide Profiles of Young Healthy Adults Acquired Using MALDI-TOF MS

A cross-sectional observational study was conducted to evaluate the inter-individual variation in the MALDI-TOF MS peptide profiles of unstimulated whole saliva in a population of 268 systemically healthy adults aged 18-30 yr (150 males and 118 females) with no apparent caries lesions or periodontal disease. Using Spectral Clustering, four subgroups of individuals were identified within the study population. These subgroups were delimited by the pattern of variation in 9 peaks detected in the 2-15 kDa m/z range. An Unsupervised Feature Selection algorithm showed that P-C peptide, a 44 residue-long salivary acidic proline-rich protein, and three of its fragments (Fr. 1-25, Fr. 15-35 and Fr. 15-44) play a central role in delimiting the subgroups. Significant differences were found in the salivary biochemistry of the subgroups with regard to lysozyme and chitinase, two enzymes that are part of the salivary innate defense system (p < 0.001). These results suggest that MALDI-TOF MS salivary peptide profiles may relate information on the underlying state of the oral ecosystem and may provide a useful reference for salivary disease biomarker discovery studies.

Lactobacillus-dominated cervicovaginal microbiota associated with reduced HIV/STI prevalence and genital HIV viral load in African women

Cervicovaginal microbiota not dominated by lactobacilli may facilitate transmission of HIV and other sexually transmitted infections (STIs), as well as miscarriages, preterm births and sepsis in pregnant women. However, little is known about the exact nature of the microbiological changes that cause these adverse outcomes. In this study, cervical samples of 174 Rwandan female sex workers were analyzed cross-sectionally using a phylogenetic microarray. Furthermore, HIV-1 RNA concentrations were measured in cervicovaginal lavages of 58 HIV-positive women among them. We identified six microbiome clusters, representing a gradient from low semi-quantitative abundance and diversity dominated by Lactobacillus crispatus (cluster R-I, with R denoting 'Rwanda') and L. iners (R-II) to intermediate (R-V) and high abundance and diversity (R-III, R-IV and R-VI) dominated by a mixture of anaerobes, including Gardnerella, Atopobium and Prevotella species. Women in cluster R-I were less likely to have HIV (P=0.03), herpes simplex virus type 2 (HSV-2; P<0.01), and high-risk human papillomavirus (HPV; P<0.01) and had no bacterial STIs (P=0.15). Statistically significant trends in prevalence of viral STIs were found from low prevalence in cluster R-I, to higher prevalence in clusters R-II and R-V, and highest prevalence in clusters R-III/R-IV/R-VI. Furthermore, only 10% of HIV-positive women in clusters R-I/R-II, compared with 40% in cluster R-V, and 42% in clusters R-III/R-IV/R-VI had detectable cervicovaginal HIV-1 RNA (Ptrend=0.03). We conclude that L. crispatus-dominated, and to a lesser extent L. iners-dominated, cervicovaginal microbiota are associated with a lower prevalence of HIV/STIs and a lower likelihood of genital HIV-1 RNA shedding.