mb-PHENIX: diffusion and supervised uniform manifold approximation for denoizing microbiota data.

MOTIVATION: Microbiota data encounters challenges arising from technical noise and the curse of dimensionality, which affect the reliability of scientific findings. Furthermore, abundance matrices exhibit a zero-inflated distribution due to biological and technical influences. Consequently, there is a growing demand for advanced algorithms that can effectively recover missing taxa while also considering the preservation of data structure. RESULTS: We present mb-PHENIX, an open-source algorithm developed in Python that recovers taxa abundances from the noisy and sparse microbiota data. Our method infers the missing information of count matrix (in 16S microbiota and shotgun studies) by applying imputation via diffusion with supervised Uniform Manifold Approximation Projection (sUMAP) space as initialization. Our hybrid machine learning approach allows to denoise microbiota data, revealing differential abundance microbes among study groups where traditional abundance analysis fails. AVAILABILITY AND IMPLEMENTATION: The mb-PHENIX algorithm is available at https://github.com/resendislab/mb-PHENIX. An easy-to-use implementation is available on Google Colab (see GitHub).

Machine Learning and COVID-19: Lessons from SARS-CoV-2.

Currently, methods in machine learning have opened a significant number of applications to construct classifiers with capacities to recognize, identify, and interpret patterns hidden in massive amounts of data. This technology has been used to solve a variety of social and health issues against coronavirus disease 2019 (COVID-19). In this chapter, we present some supervised and unsupervised machine learning techniques that have contributed in three aspects to supplying information to health authorities and diminishing the deadly effects of the current worldwide outbreak on the population. First is the identification and construction of powerful classifiers capable of predicting severe, moderate, or asymptomatic responses in COVID-19 patients starting from clinical or high-throughput technologies. Second is the identification of groups of patients with similar physiological responses to improve the triage classification and inform treatments. The final aspect is the combination of machine learning methods and schemes from systems biology to link associative studies with mechanistic frameworks. This chapter aims to discuss some practical applications in the use of machine learning techniques to handle data coming from social behavior and high-throughput technologies, associated with COVID-19 evolution.

Uncoding the interdependency of tumor microenvironment and macrophage polarization: insights from a continuous network approach.

The balance between pro- and anti-inflammatory immune system responses is crucial to preventing complex diseases like cancer. Macrophages are essential immune cells that contribute to this balance constrained by the local signaling profile of the tumor microenvironment. To understand how pro- and anti-inflammatory unbalance emerges in cancer, we developed a theoretical analysis of macrophage differentiation that is derived from activated monocytes circulating in the blood. Once recruited to the site of inflammation, monocytes can be polarized based on the specific interleukins and chemokines in the microenvironment. To quantify this process, we used a previous regulatory network reconstructed by our group and transformed Boolean Network attractors of macrophage polarization to an ODE scheme, it enables us to quantify the activation of their genes in a continuous fashion. The transformation was developed using the interaction rules with a fuzzy logic approach. By implementing this approach, we analyzed different aspects that cannot be visualized in the Boolean setting. For example, this approach allows us to explore the dynamic behavior at different concentrations of cytokines and transcription factors in the microenvironment. One important aspect to assess is the evaluation of the transitions between phenotypes, some of them characterized by an abrupt or a gradual transition depending on specific concentrations of exogenous cytokines in the tumor microenvironment. For instance, IL-10 can induce a hybrid state that transits between an M2c and an M2b macrophage. Interferon- γ can induce a hybrid between M1 and M1a macrophage. We further demonstrated the plasticity of macrophages based on a combination of cytokines and the existence of hybrid phenotypes or partial polarization. This mathematical model allows us to unravel the patterns of macrophage differentiation based on the competition of expression of transcriptional factors. Finally, we survey how macrophages may respond to a continuously changing immunological response in a tumor microenvironment.

Dysbiosis signatures of gut microbiota and the progression of type 2 diabetes: a machine learning approach in a Mexican cohort.

Introduction: The gut microbiota (GM) dysbiosis is one of the causal factors for the progression of different chronic metabolic diseases, including type 2 diabetes mellitus (T2D). Understanding the basis that laid this association may lead to developing new therapeutic strategies for preventing and treating T2D, such as probiotics, prebiotics, and fecal microbiota transplants. It may also help identify potential early detection biomarkers and develop personalized interventions based on an individual's gut microbiota profile. Here, we explore how supervised Machine Learning (ML) methods help to distinguish taxa for individuals with prediabetes (prediabetes) or T2D. Methods: To this aim, we analyzed the GM profile (16s rRNA gene sequencing) in a cohort of 410 Mexican naïve patients stratified into normoglycemic, prediabetes, and T2D individuals. Then, we compared six different ML algorithms and found that Random Forest had the highest predictive performance in classifying T2D and prediabetes patients versus controls. Results: We identified a set of taxa for predicting patients with T2D compared to normoglycemic individuals, including Allisonella, Slackia, Ruminococus_2, Megaspgaera, Escherichia/Shigella, and Prevotella, among them. Besides, we concluded that Anaerostipes, Intestinibacter, Prevotella_9, Blautia, Granulicatella, and Veillonella were the relevant genus in patients with prediabetes compared to normoglycemic subjects. Discussion: These findings allow us to postulate that GM is a distinctive signature in prediabetes and T2D patients during the development and progression of the disease. Our study highlights the role of GM and opens a window toward the rational design of new preventive and personalized strategies against the control of this disease.

Spermiogenesis alterations in the absence of CTCF revealed by single cell RNA sequencing.

CTCF is an architectonic protein that organizes the genome inside the nucleus in almost all eukaryotic cells. There is evidence that CTCF plays a critical role during spermatogenesis as its depletion produces abnormal sperm and infertility. However, defects produced by its depletion throughout spermatogenesis have not been fully characterized. In this work, we performed single cell RNA sequencing in spermatogenic cells with and without CTCF. We uncovered defects in transcriptional programs that explain the severity of the damage in the produced sperm. In the early stages of spermatogenesis, transcriptional alterations are mild. As germ cells go through the specialization stage or spermiogenesis, transcriptional profiles become more altered. We found morphology defects in spermatids that support the alterations in their transcriptional profiles. Altogether, our study sheds light on the contribution of CTCF to the phenotype of male gametes and provides a fundamental description of its role at different stages of spermiogenesis.

Boolean modeling reveals that cyclic attractors in macrophage polarization serve as reservoirs of states to balance external perturbations from the tumor microenvironment.

Cyclic attractors generated from Boolean models may explain the adaptability of a cell in response to a dynamical complex tumor microenvironment. In contrast to this idea, we postulate that cyclic attractors in certain cases could be a systemic mechanism to face the perturbations coming from the environment. To justify our conjecture, we present a dynamic analysis of a highly curated transcriptional regulatory network of macrophages constrained into a cancer microenvironment. We observed that when M1-associated transcription factors (STAT1 or NF-κB) are perturbed and the microenvironment balances to a hyper-inflammation condition, cycle attractors activate genes whose signals counteract this effect implicated in tissue damage. The same behavior happens when the M2-associated transcription factors are disturbed (STAT3 or STAT6); cycle attractors will prevent a hyper-regulation scenario implicated in providing a suitable environment for tumor growth. Therefore, here we propose that cyclic macrophage phenotypes can serve as a reservoir for balancing the phenotypes when a specific phenotype-based transcription factor is perturbed in the regulatory network of macrophages. We consider that cyclic attractors should not be simply ignored, but it is necessary to carefully evaluate their biological importance. In this work, we suggest one conjecture: the cyclic attractors can serve as a reservoir to balance the inflammatory/regulatory response of the network under external perturbations.

Transcriptional and Microenvironmental Landscape of Macrophage Transition in Cancer: A Boolean Analysis.

The balance between pro- and anti-inflammatory immune system responses is crucial to face and counteract complex diseases such as cancer. Macrophages are an essential population that contributes to this balance in collusion with the local tumor microenvironment. Cancer cells evade the attack of macrophages by liberating cytokines and enhancing the transition to the M2 phenotype with pro-tumoral functions. Despite this pernicious effect on immune systems, the M1 phenotype still exists in the environment and can eliminate tumor cells by liberating cytokines that recruit and activate the cytotoxic actions of TH1 effector cells. Here, we used a Boolean modeling approach to understand how the tumor microenvironment shapes macrophage behavior to enhance pro-tumoral functions. Our network reconstruction integrates experimental data and public information that let us study the polarization from monocytes to M1, M2a, M2b, M2c, and M2d subphenotypes. To analyze the dynamics of our model, we modeled macrophage polarization in different conditions and perturbations. Notably, our study identified new hybrid cell populations, undescribed before. Based on the in vivo macrophage behavior, we explained the hybrid macrophages' role in the tumor microenvironment. The in silico model allowed us to postulate transcriptional factors that maintain the balance between macrophages with anti- and pro-tumoral functions. In our pursuit to maintain the balance of macrophage phenotypes to eliminate malignant tumor cells, we emulated a theoretical genetically modified macrophage by modifying the activation of NFκB and a loss of function in HIF1-α and discussed their phenotype implications. Overall, our theoretical approach is as a guide to design new experiments for unraveling the principles of the dual host-protective or -harmful antagonistic roles of transitional macrophages in tumor immunoediting and cancer cell fate decisions.

On Deep Landscape Exploration of COVID-19 Patients Cells and Severity Markers.

COVID-19 is a disease with a spectrum of clinical responses ranging from moderate to critical. To study and control its effects, a large number of researchers are focused on two substantial aims. On the one hand, the discovery of diverse biomarkers to classify and potentially anticipate the disease severity of patients. These biomarkers could serve as a medical criterion to prioritize attention to those patients with higher prone to severe responses. On the other hand, understanding how the immune system orchestrates its responses in this spectrum of disease severities is a fundamental issue required to design new and optimized therapeutic strategies. In this work, using single-cell RNAseq of bronchoalveolar lavage fluid of nine patients with COVID-19 and three healthy controls, we contribute to both aspects. First, we presented computational supervised machine-learning models with high accuracy in classifying the disease severity (moderate and severe) in patients with COVID-19 starting from single-cell data from bronchoalveolar lavage fluid. Second, we identified regulatory mechanisms from the heterogeneous cell populations in the lungs microenvironment that correlated with different clinical responses. Given the results, patients with moderate COVID-19 symptoms showed an activation/inactivation profile for their analyzed cells leading to a sequential and innocuous immune response. In comparison, severe patients might be promoting cytotoxic and pro-inflammatory responses in a systemic fashion involving epithelial and immune cells without the possibility to develop viral clearance and immune memory. Consequently, we present an in-depth landscape analysis of how transcriptional factors and pathways from these heterogeneous populations can regulate their expression to promote or restrain an effective immune response directly linked to the patients prognosis.

MicroRNAs Regulate Metabolic Phenotypes During Multicellular Tumor Spheroids Progression.

During tumor progression, cancer cells rewire their metabolism to face their bioenergetic demands. In recent years, microRNAs (miRNAs) have emerged as regulatory elements that inhibit the translation and stability of crucial mRNAs, some of them causing direct metabolic alterations in cancer. In this study, we investigated the relationship between miRNAs and their targets mRNAs that control metabolism, and how this fine-tuned regulation is diversified depending on the tumor stage. To do so, we implemented a paired analysis of RNA-seq and small RNA-seq in a breast cancer cell line (MCF7). The cell line was cultured in multicellular tumor spheroid (MCTS) and monoculture conditions. For MCTS, we selected two-time points during their development to recapitulate a proliferative and quiescent stage and contrast their miRNA and mRNA expression patterns associated with metabolism. As a result, we identified a set of new direct putative regulatory interactions between miRNAs and metabolic mRNAs representative for proliferative and quiescent stages. Notably, our study allows us to suggest that miR-3143 regulates the carbon metabolism by targeting hexokinase-2. Also, we found that the overexpression of several miRNAs could directly overturn the expression of mRNAs that control glycerophospholipid and N-Glycan metabolism. While this set of miRNAs downregulates their expression in the quiescent stage, the same set is upregulated in proliferative stages. This last finding suggests an additional metabolic switch of the above mentioned metabolic pathways between the quiescent and proliferative stages. Our results contribute to a better understanding of how miRNAs modulate the metabolic landscape in breast cancer MCTS, which eventually will help to design new strategies to mitigate cancer phenotype.

Unveiling functional heterogeneity in breast cancer multicellular tumor spheroids through single-cell RNA-seq.

Heterogeneity is an intrinsic characteristic of cancer. Even in isogenic tumors, cell populations exhibit differential cellular programs that overall supply malignancy and decrease treatment efficiency. In this study, we investigated the functional relationship among cell subtypes and how this interdependency can promote tumor development in a cancer cell line. To do so, we performed single-cell RNA-seq of MCF7 Multicellular Tumor Spheroids as a tumor model. Analysis of single-cell transcriptomes at two-time points of the spheroid growth, allowed us to dissect their functional relationship. As a result, three major robust cellular clusters, with a non-redundant complementary composition, were found. Meanwhile, one cluster promotes proliferation, others mainly activate mechanisms to invade other tissues and serve as a reservoir population conserved over time. Our results provide evidence to see cancer as a systemic unit that has cell populations with task stratification with the ultimate goal of preserving the hallmarks in tumors.

On Information Extraction and Decoding Mechanisms Improved by Noisy Amplification in Signaling Pathways.

The cells need to process information about extracellular stimuli. They encode, transmit and decode the information to elicit an appropriate response. Studies aimed at understanding how such information is decoded in the signaling pathways to generate a specific cellular response have become essential. Eukaryotic cells decode information through two different mechanisms: the feed-forward loop and the promoter affinity. Here, we investigate how these two mechanisms improve information transmission. A detailed comparison is made between the stochastic model of the MAPK/ERK pathway and a stochastic minimal decoding model. The maximal amount of transmittable information was computed. The results suggest that the decoding mechanism of the MAPK/ERK pathway improve the channel capacity because it behaves as a noisy amplifier. We show a positive dependence between the noisy amplification and the amount of information extracted. Additionally, we show that the extrinsic noise can be tuned to improve information transmission. This investigation has revealed that the feed-forward loop and the promoter affinity motifs extract information thanks to processes of amplification and noise addition. Moreover, the channel capacity is enhanced when both decoding mechanisms are coupled. Altogether, these findings suggest novel characteristics in how decoding mechanisms improve information transmission.