Gut bacteria are closely associated with the development of atopic dermatitis (AD) due to their immunoregulatory function. Indole derivatives, produced by gut bacteria metabolizing tryptophan, are ligands to activate the aryl hydrocarbon receptor (AHR), which plays a critical role in attenuating AD symptoms. Limosilactobacillus reuteri, a producer of indole derivatives, regulates mucosal immunity via activating the AHR signaling pathway. However, the effective substance and mechanism of L. reuteri in the amelioration of AD remain to be elucidated. In this research, we found that L. reuteri DYNDL22M62 significantly improved AD-like symptoms in mice by suppressing IgE levels and the expressions of thymic stromal lymphopoietin (TSLP), IL-4, and IL-5. L. reuteri DYNDL22M62 induced an increase in the production of indole lactic acid (ILA) and indole propionic acid (IPA) via targeted tryptophan metabolic analysis and the expression of AHR in mice. Furthermore, L. reuteri DYNDL22M62 increased the proportions of Romboutsia and Ruminococcaceae NK4A214 group, which were positively related to ILA, but decreased Dubosiella, which was negatively related to IPA. Collectively, L. reuteri DYNDL22M62 with the role of modulating gut bacteria and the production of indole derivatives may attenuate AD via activating AHR in mice.

Batch processes are commonly involved by a succession of working phases with implicit non-Gaussian behaviors. Besides, in most cases, batch-to-batch processes also show similar but yet not identical running trajectory variations. To deal with these issues, this paper introduces a systematic analysis flowchart based on local outlier factor (LOF) for monitoring multiphase batch processes. First, a step-wise phase dividing algorithm is proposed with LOF to conduct phase dividing for a better understanding of batch process. Afterward, we develop the multiphase LOF for similar batch data modeling and then fault detection. The fault isolation method is proposed, where variable contributions with LOFs are induced, also with the analysis of isolability. The developed method is validated on a numerical example and the fed-batch fermentation benchmark process, both of which are compared with the multiphase principal component analysis method. Results demonstrate the feasibility and superiority of the proposed method.

This work considers industrial process monitoring using a variational autoencoder (VAE). As a powerful deep generative model, the variational autoencoder and its variants have become popular for process monitoring. However, its monitoring ability, especially its fault diagnosis ability, has not been well investigated. In this paper, the process modeling and monitoring capabilities of several VAE variants are comprehensively studied. First, fault detection schemes are defined in three distinct ways, considering latent, residual, and the combined domains. Afterwards, to conduct the fault diagnosis, we first define the deep contribution plot, and then a deep reconstruction-based contribution diagram is proposed for deep domains under the fault propagation mechanism. In a case study, the performance of the process monitoring capability of four deep VAE models, namely, the static VAE model, the dynamic VAE model, and the recurrent VAE models (LSTM-VAE and GRU-VAE), has been comparatively evaluated on the industrial benchmark Tennessee Eastman process. Results show that recurrent VAEs with a deep reconstruction-based diagnosis mechanism are recommended for industrial process monitoring tasks.

Background: The gut microbiome has proven to be an important factor affecting obesity; however, it remains a challenge to identify consistent biomarkers across geographic locations and perform precisely targeted modulation for obese individuals. Results: This study proposed a systematic machine learning framework and applied it to 870 human stool metagenomes across five countries to obtain comprehensive regional shared biomarkers and conduct a personalized modulation analysis. In our pipeline, a heterogeneous ensemble feature selection diagram is first developed to determine an optimal subset of biomarkers through the aggregation of multiple techniques. Subsequently, a deep reinforcement learning method was established to alter the targeted composition to the desired healthy target. In this manner, we can realize personalized modulation by counterfactual inference. Consequently, a total of 42 species were identified as regional shared biomarkers, and they showed good performance in distinguishing obese people from the healthy group (area under curve (AUC) =0.85) when demonstrated on validation datasets. In addition, by pooling all counterfactual explanations, we found that Akkermansia muciniphila, Faecalibacterium prausnitzii, Prevotella copri, Bacteroides dorei, Bacteroides eggerthii, Alistipes finegoldii, Alistipes shahii, Eubacterium sp. _CAG_180, and Roseburia hominis may be potential broad-spectrum targets with consistent modulation in the multi-regional obese population. Conclusions: This article shows that based on our proposed machine-learning framework, we can obtain more comprehensive and accurate biomarkers and provide modulation analysis for the obese population. Moreover, our machine-learning framework will also be very useful for other researchers to further obtain biomarkers and perform counterfactual modulation analysis in different diseases.

A growing number of studies show that the human microbiome plays a vital role in human health and can be a crucial factor in predicting certain human diseases. However, microbiome data are often characterized by the limited samples and high-dimensional features, which pose a great challenge for machine learning methods. Therefore, this paper proposes a novel ensemble deep learning disease prediction method that combines unsupervised and supervised learning paradigms. First, unsupervised deep learning methods are used to learn the potential representation of the sample. Afterwards, the disease scoring strategy is developed based on the deep representations as the informative features for ensemble analysis. To ensure the optimal ensemble, a score selection mechanism is constructed, and performance boosting features are engaged with the original sample. Finally, the composite features are trained with gradient boosting classifier for health status decision. For case study, the ensemble deep learning flowchart has been demonstrated on six public datasets extracted from the human microbiome profiling. The results show that compared with the existing algorithms, our framework achieves better performance on disease prediction.