Machine learning-based investigation of the relationship between gut microbiome and obesity status

Gut bacteriome and mood disorders in women with PCOS

STUDY QUESTION

How does the gut bacteriome differ based on mood disorders (MDs) in women with polycystic ovary syndrome (PCOS), and how can the gut bacteriome contribute to the associations between these two conditions?

SUMMARY ANSWER

Women with PCOS who also have MDs exhibited a distinct gut bacteriome with reduced alpha diversity and a significantly lower abundance of Butyricicoccus compared to women with PCOS but without MDs.

WHAT IS KNOWN ALREADY

Women with PCOS have a 4- to 5-fold higher risk of having MDs compared to women without PCOS. The gut bacteriome has been suggested to influence the pathophysiology of both PCOS and MDs.

STUDY DESIGN, SIZE, DURATION

This population-based cohort study was derived from the Northern Finland Birth Cohort 1966 (NFBC1966), which includes all women born in Northern Finland in 1966. Women with PCOS who donated a stool sample at age 46 years (n = 102) and two BMI-matched controls for each case (n = 205), who also responded properly to the MD criteria scales, were included.

PARTICIPANTS/MATERIALS, SETTING, METHODS

A total of 102 women with PCOS and 205 age- and BMI-matched women without PCOS were included. Based on the validated MD criteria, the subjects were categorized into MD or no-MD groups, resulting in the following subgroups: PCOS no-MD (n = 84), PCOS MD (n = 18), control no-MD (n = 180), and control MD (n = 25). Clinical characteristics were assessed at age 31 years and age 46 years, and stool samples were collected from the women at age 46 years, followed by the gut bacteriome analysis using 16 s rRNA sequencing. Alpha diversity was assessed using observed features and Shannon's index, with a focus on genera, and beta diversity was characterized using principal components analysis (PCA) with Bray-Curtis Dissimilarity at the genus level. Associations between the gut bacteriome and PCOS-related clinical features were explored by Spearman's correlation coefficient. A P-value for multiple testing was adjusted with the Benjamini-Hochberg false discovery rate (FDR) method.

MAIN RESULTS AND THE ROLE OF CHANCE

We observed changes in the gut bacteriome associated with MDs, irrespective of whether the women also had PCOS. Similarly, PCOS MD cases showed a lower alpha diversity (Observed feature, PCOS no-MD, median 272; PCOS MD, median 208, FDR = 0.01; Shannon, PCOS no-MD, median 5.95; PCOS MD, median 5.57, FDR = 0.01) but also a lower abundance of Butyricicoccus (log-fold changeAnalysis of Compositions of Microbiomes with Bias Correction (ANCOM-BC)=-0.90, FDRANCOM-BC=0.04) compared to PCOS no-MD cases. In contrast, in the controls, the gut bacteriome did not differ based on MDs. Furthermore, in the PCOS group, Sutterella showed positive correlations with PCOS-related clinical parameters linked to obesity (BMI, r2=0.31, FDR = 0.01; waist circumference, r2=0.29, FDR = 0.02), glucose metabolism (fasting glucose, r2=0.46, FDR < 0.001; fasting insulin, r2=0.24, FDR = 0.05), and gut barrier integrity (zonulin, r2=0.25, FDR = 0.03).

LIMITATIONS, REASONS FOR CAUTION

Although this was the first study to assess the link between the gut bacteriome and MDs in PCOS and included the largest PCOS dataset for the gut microbiome analysis, the number of subjects stratified by the presence of MDs was limited when contrasted with previous studies that focused on MDs in a non-selected population.

WIDER IMPLICATIONS OF THE FINDINGS

The main finding is that gut bacteriome is associated with MDs irrespective of the PCOS status, but PCOS may also modulate further the connection between the gut bacteriome and MDs.

STUDY FUNDING/COMPETING INTEREST(S)

This research was funded by the European Union's Horizon 2020 Research and Innovation Programme under the Marie Sklodowska-Curie Grant Agreement (MATER, No. 813707), the Academy of Finland (project grants 315921, 321763, 336449), the Sigrid Jusélius Foundation, Novo Nordisk Foundation (NNF21OC0070372), grant numbers PID2021-12728OB-100 (Endo-Map) and CNS2022-135999 (ROSY) funded by MCIN/AEI/10.13039/501100011033 and ERFD A Way of Making Europe. The study was also supported by EU QLG1-CT-2000-01643 (EUROBLCS) (E51560), NorFA (731, 20056, 30167), USA/NIH 2000 G DF682 (50945), the Estonian Research Council (PRG1076, PRG1414), EMBO Installation (3573), and Horizon 2020 Innovation Grant (ERIN, No. EU952516). The funders did not participate in any process of the study. We have no conflicts of interest to declare.

TRIAL REGISTRATION NUMBER

N/A.

A review of machine learning methods for cancer characterization from microbiome data

Recent studies have shown that the microbiome can impact cancer development, progression, and response to therapies suggesting microbiome-based approaches for cancer characterization. As cancer-related signatures are complex and implicate many taxa, their discovery often requires Machine Learning approaches. This review discusses Machine Learning methods for cancer characterization from microbiome data. It focuses on the implications of choices undertaken during sample collection, feature selection and pre-processing. It also discusses ML model selection, guiding how to choose an ML model, and model validation. Finally, it enumerates current limitations and how these may be surpassed. Proposed methods, often based on Random Forests, show promising results, however insufficient for widespread clinical usage. Studies often report conflicting results mainly due to ML models with poor generalizability. We expect that evaluating models with expanded, hold-out datasets, removing technical artifacts, exploring representations of the microbiome other than taxonomical profiles, leveraging advances in deep learning, and developing ML models better adapted to the characteristics of microbiome data will improve the performance and generalizability of models and enable their usage in the clinic.

Gut microbial interactions based on network construction and bacterial pairwise cultivation

Association networks are widely applied for the prediction of bacterial interactions in studies of human gut microbiomes. However, the experimental validation of the predicted interactions is challenging due to the complexity of gut microbiomes and the limited number of cultivated bacteria. In this study, we addressed this challenge by integrating in vitro time series network (TSN) associations and co-cultivation of TSN taxon pairs. Fecal samples were collected and used for cultivation and enrichment of gut microbiome on YCFA agar plates for 13 days. Enriched cells were harvested for DNA extraction and metagenomic sequencing. A total of 198 metagenome-assembled genomes (MAGs) were recovered. Temporal dynamics of bacteria growing on the YCFA agar were used to infer microbial association networks. To experimentally validate the interactions of taxon pairs in networks, we selected 24 and 19 bacterial strains from this study and from the previously established human gut microbial biobank, respectively, for pairwise co-cultures. The co-culture experiments revealed that most of the interactions between taxa in networks were identified as neutralism (51.67%), followed by commensalism (21.67%), amensalism (18.33%), competition (5%) and exploitation (3.33%). Genome-centric analysis further revealed that the commensal gut bacteria (helpers and beneficiaries) might interact with each other via the exchanges of amino acids with high biosynthetic costs, short-chain fatty acids, and/or vitamins. We also validated 12 beneficiaries by adding 16 additives into the basic YCFA medium and found that the growth of 66.7% of these strains was significantly promoted. This approach provides new insights into the gut microbiome complexity and microbial interactions in association networks. Our work highlights that the positive relationships in gut microbial communities tend to be overestimated, and that amino acids, short-chain fatty acids, and vitamins are contributed to the positive relationships.

Integrative metagenomic analysis reveals distinct gut microbial signatures related to obesity

Obesity is a metabolic disorder closely associated with profound alterations in gut microbial composition. However, the dynamics of species composition and functional changes in the gut microbiome in obesity remain to be comprehensively investigated. In this study, we conducted a meta-analysis of metagenomic sequencing data from both obese and non-obese individuals across multiple cohorts, totaling 1351 fecal metagenomes. Our results demonstrate a significant decrease in both the richness and diversity of the gut bacteriome and virome in obese patients. We identified 38 bacterial species including Eubacterium sp. CAG:274, Ruminococcus gnavus, Eubacterium eligens and Akkermansia muciniphila, and 1 archaeal species, Methanobrevibacter smithii, that were significantly altered in obesity. Additionally, we observed altered abundance of five viral families: Mesyanzhinovviridae, Chaseviridae, Salasmaviridae, Drexlerviridae, and Casjensviridae. Functional analysis of the gut microbiome indicated distinct signatures associated to obesity and identified Ruminococcus gnavus as the primary driver for function enrichment in obesity, and Methanobrevibacter smithii, Akkermansia muciniphila, Ruminococcus bicirculans, and Eubacterium siraeum as functional drivers in the healthy control group. Additionally, our results suggest that antibiotic resistance genes and bacterial virulence factors may influence the development of obesity. Finally, we demonstrated that gut vOTUs achieved a diagnostic accuracy with an optimal area under the curve of 0.766 for distinguishing obesity from healthy controls. Our findings offer comprehensive and generalizable insights into the gut bacteriome and virome features associated with obesity, with the potential to guide the development of microbiome-based diagnostics.

1-Deoxynojirimycin containing Morus alba leaf-based food modulates the gut microbiome and expression of genes related to obesity

BACKGROUND

Obesity is a serious disease with an alarmingly high incidence that can lead to other complications in both humans and dogs. Similar to humans, obesity can cause metabolic diseases such as diabetes in dogs. Natural products may be the preferred intervention for metabolic diseases such as obesity. The compound 1-deoxynojirimycin, present in Morus leaves and other sources has antiobesity effects. The possible antiobesity effect of 1-deoxynojirimycin containing Morus alba leaf-based food was studied in healthy companion dogs (n = 46) visiting the veterinary clinic without a history of diseases. Body weight, body condition score (BCS), blood-related parameters, and other vital parameters of the dogs were studied. Whole-transcriptome of blood and gut microbiome analysis was also carried out to investigate the possible mechanisms of action and role of changes in the gut microbiome due to treatment.

RESULTS

After 90 days of treatment, a significant antiobesity effect of the treatment food was observed through the reduction of weight, BCS, and blood-related parameters. A whole-transcriptome study revealed differentially expressed target genes important in obesity and diabetes-related pathways such as MLXIPL, CREB3L1, EGR1, ACTA2, SERPINE1, NOTCH3, and CXCL8. Gut microbiome analysis also revealed a significant difference in alpha and beta-diversity parameters in the treatment group. Similarly, the microbiota known for their health-promoting effects such as Lactobacillus ruminis, and Weissella hellenica were abundant (increased) in the treatment group. The predicted functional pathways related to obesity were also differentially abundant between groups.

CONCLUSIONS

1-Deoxynojirimycin-containing treatment food have been shown to significantly improve obesity. The identified genes, pathways, and gut microbiome-related results may be pursued in further studies to develop 1-deoxynojirimycin-based products as candidates against obesity.

A polysaccharide from salviae miltiorrhizae radix inhibits weight gain of mice with high-fat diet via modulating intestinal bacteria

BACKGROUND

Obesity, a global chronic disease, has been recognized as a severe risk to health. In our study, a novel polysaccharide named ARS was isolated and purified from aerial part of salviae miltiorrhizae radix. Our aim is to investigate the weight-reducing effect of a polysaccharide from salviae miltiorrhizae radix on mice fed a high-fat diet.

RESULTS

The novel polysaccharide ARS mainly consisted of glucose and galactose with a molar ratio of 0.59:1.00. We found that treatment with ARS could inhibit weight gain of mice fed a high-fat diet via modulating the intestinal bacteria. Moreover, we surveyed its mechanism in mice, and the gut microbiota sequencing results demonstrated that ARS can reverse or resist high-fat-diet-induced significant weight gain or obesity by increasing the diversity of gut microbiota and optimizing the ratio of Firmicutes to Bacteroidetes. Phylum and species analysis of gut microbiota demonstrated that obesity caused by a high-fat diet was accompanied by significant changes in the microbial communities, but ARS could reverse the disturbance of gut microbiota induced by the high-fat diet to maintain homeostasis.

CONCLUSION

Overall, our findings suggested a new function of ARS in regulating gut microbiota, which provides a theoretical basis for the development of high-quality ARS functional foods and the application of dietary supplements. © 2023 Society of Chemical Industry.

Overview of data preprocessing for machine learning applications in human microbiome research

Although metagenomic sequencing is now the preferred technique to study microbiome-host interactions, analyzing and interpreting microbiome sequencing data presents challenges primarily attributed to the statistical specificities of the data (e.g., sparse, over-dispersed, compositional, inter-variable dependency). This mini review explores preprocessing and transformation methods applied in recent human microbiome studies to address microbiome data analysis challenges. Our results indicate a limited adoption of transformation methods targeting the statistical characteristics of microbiome sequencing data. Instead, there is a prevalent usage of relative and normalization-based transformations that do not specifically account for the specific attributes of microbiome data. The information on preprocessing and transformations applied to the data before analysis was incomplete or missing in many publications, leading to reproducibility concerns, comparability issues, and questionable results. We hope this mini review will provide researchers and newcomers to the field of human microbiome research with an up-to-date point of reference for various data transformation tools and assist them in choosing the most suitable transformation method based on their research questions, objectives, and data characteristics.

From-Toilet-to-Freezer: A Review on Requirements for an Automatic Protocol to Collect and Store Human Fecal Samples for Research Purposes

The composition, viability and metabolic functionality of intestinal microbiota play an important role in human health and disease. Studies on intestinal microbiota are often based on fecal samples, because these can be sampled in a non-invasive way, although procedures for sampling, processing and storage vary. This review presents factors to consider when developing an automated protocol for sampling, processing and storing fecal samples: donor inclusion criteria, urine-feces separation in smart toilets, homogenization, aliquoting, usage or type of buffer to dissolve and store fecal material, temperature and time for processing and storage and quality control. The lack of standardization and low-throughput of state-of-the-art fecal collection procedures promote a more automated protocol. Based on this review, an automated protocol is proposed. Fecal samples should be collected and immediately processed under anaerobic conditions at either room temperature (RT) for a maximum of 4 h or at 4 °C for no more than 24 h. Upon homogenization, preferably in the absence of added solvent to allow addition of a buffer of choice at a later stage, aliquots obtained should be stored at either -20 °C for up to a few months or -80 °C for a longer period-up to 2 years. Protocols for quality control should characterize microbial composition and viability as well as metabolic functionality.

Machine learning insights concerning inflammatory and liver-related risk comorbidities in non-communicable and viral diseases

The liver is a key organ involved in a wide range of functions, whose damage can lead to chronic liver disease (CLD). CLD accounts for more than two million deaths worldwide, becoming a social and economic burden for most countries. Among the different factors that can cause CLD, alcohol abuse, viruses, drug treatments, and unhealthy dietary patterns top the list. These conditions prompt and perpetuate an inflammatory environment and oxidative stress imbalance that favor the development of hepatic fibrogenesis. High stages of fibrosis can eventually lead to cirrhosis or hepatocellular carcinoma (HCC). Despite the advances achieved in this field, new approaches are needed for the prevention, diagnosis, treatment, and prognosis of CLD. In this context, the scientific com-munity is using machine learning (ML) algorithms to integrate and process vast amounts of data with unprecedented performance. ML techniques allow the integration of anthropometric, genetic, clinical, biochemical, dietary, lifestyle and omics data, giving new insights to tackle CLD and bringing personalized medicine a step closer. This review summarizes the investigations where ML techniques have been applied to study new approaches that could be used in inflammatory-related, hepatitis viruses-induced, and coronavirus disease 2019-induced liver damage and enlighten the factors involved in CLD development.

Influence of Adiposity on the Gut Microbiota Composition of Arab Women: A Case-Control Study

Recent evidence has suggested that the gut microbiota is a possible risk factor for obesity. However, limited evidence is available on the association between the gut microbiota composition and obesity markers in the Middle-Eastern region. We aimed to investigate the association between gut microbiota and obesity markers in a case-control study including 92 Saudi women aged 18-25 years, including participants with obesity (case, n = 44) and with normal weight (control, n = 48). Anthropometric, body composition, and biochemical data were collected. The whole-genome shotgun technique was used to analyze the gut microbiota. The Shannon alpha and Bray-Curtis beta diversity were determined. The microbial alpha diversity was significantly associated with only the waist-to-hip ratio (WHR) (p-value = 0.04), while the microbial beta diversity was significantly associated with body mass index (p-value = 0.048), %body fat (p-value = 0.018), and WHR (p-value = 0.050). Specific bacteria at different taxonomic levels, such as Bacteroidetes and Synergistetes, were positively associated with different obesity markers. Alistipes was higher in the control group compared with the case group. The results highlight the association of the gut microbiota with obesity and suggest that the gut microbiota of Saudi women is associated with specific obesity markers. Future studies are needed to determine the role of the identified strains in the metabolism of individuals with obesity.

BiGAMi: Bi-Objective Genetic Algorithm Fitness Function for Feature Selection on Microbiome Datasets

The relationship between the host and the microbiome, or the assemblage of microorganisms (including bacteria, archaea, fungi, and viruses), has been proven crucial for its health and disease development. The high dimensionality of microbiome datasets has often been addressed as a major difficulty for data analysis, such as the use of machine-learning (ML) and deep-learning (DL) models. Here, we present BiGAMi, a bi-objective genetic algorithm fitness function for feature selection in microbial datasets to train high-performing phenotype classifiers. The proposed fitness function allowed us to build classifiers that outperformed the baseline performance estimated by the original studies by using as few as 0.04% to 2.32% features of the original dataset. In 35 out of 42 performance comparisons between BiGAMi and other feature selection methods evaluated here (sequential forward selection, SelectKBest, and GARS), BiGAMi achieved its results by selecting 6–93% fewer features. This study showed that the application of a bi-objective GA fitness function against microbiome datasets succeeded in selecting small subsets of bacteria whose contribution to understood diseases and the host state was already experimentally proven. Applying this feature selection approach to novel diseases is expected to quickly reveal the microbes most relevant to a specific condition.