Search

MotivationMany computational tools attempt to infer gene regulatory networks (GRNs) from single-cell RNA sequencing (scRNA-seq) data. One recent advance is DeepSEM, a deep generative model generalizing the Linear Structural Equation Model (SEM) that improves benchmark performance over popular GRN inference methods. While DeepSEM is promising, its results are not stable over multiple runs. To overcome the instability and resolve dropout handling concerns, we propose GRN-VAE.ResultsGRN-VAE improves stability and efficiency while maintaining accuracy by delayed introduction of the sparse loss term. To minimize the negative impact of dropout in single-cell data, GRN-VAE trains on non-zero data. Most importantly, we introduce a novel idea, Dropout Augmentation, to improve model robustness by adding a small amount of simulated dropout to the data. GRN-VAE compares favorably to other methods on the BEELINE benchmark data sets, using several collections of “ground truth” regulatory relationships, and on a real-world data set, where it efficiently provides stable results consistent with literature-based findings.ConclusionsThe stability and robustness of GRN-VAE make it a practical and valuable addition to the toolkit for GRN inference from single-cell data. Dropout Augmentation may have wider applications beyond the GRN-inference problem.Availability and implementationSource code is available athttps://bcb.cs.tufts.edu/GRN-VAEContacthao.zhu@tufts.edu;donna.slonim@tufts.edu

SummaryMaternal immune activation is associated with adverse offspring neurodevelopmental outcomes, many of which are mediated by in utero microglial programming. Microglia remain inaccessible at birth and throughout development, thus identification of noninvasive biomarkers that can reflect fetal brain microglial programming may permit screening and intervention during critical developmental windows. Here we used lineage tracing to demonstrate the shared ontogeny between fetal brain macrophages (microglia) and fetal placental macrophages (Hofbauer cells). Single-cell RNA sequencing of murine fetal brain and placental macrophages demonstrated shared transcriptional programs. Comparison with human datasets demonstrated that placental resident macrophage signatures are highly conserved between mice and humans. Single-cell RNA-seq identified sex differences in fetal microglial and Hofbauer cell programs, and robust differences between placenta-associated maternal macrophage/monocyte (PAMM) populations in the context of a male versus a female fetus. We propose that Hofbauer cells, which are easily accessible at birth, provide novel insights into fetal brain microglial programs, potentially facilitating the early identification of offspring most vulnerable to neurodevelopmental disorders.

MotivationDrug repositioning allows expedited discovery of new applications for existing compounds, but re-screening vast compound libraries is often prohibitively expensive. “Connectivity mapping” is a process that links drugs to diseases by identifying compounds whose impact on expression in a collection of cells reverses the disease’s impact on expression in disease-relevant tissues. The high throughput LINCS project has expanded the universe of compounds and cell types for which data are available, but even with this effort, many potentially clinically useful combinations are missing. To evaluate the possibility of repurposing drugs this way despite missing data, we compared collaborative filtering with either neighborhood-based or SVD imputation methods to two naive approaches via cross-validation.ResultsMethods were evaluated for their ability to predict drug connectivity despite missing data. Predictions improved when cell type was taken into account. Neighborhood-based collaborative filtering was the most successful method, with the best improvements in non-immortalized primary cells. We also explored which classes of compounds are most and least reliant on cell type for accurate imputation, and we identified connections between related compounds even when many were not measured in the relevant cells. We conclude that even for cells in which drug responses have not been fully characterized, it is possible to identify unassayed drugs that reverse in those cells the expression signatures observed in disease.Contactdonna.slonim@tufts.edu

Identification of functional pathways mediating molecular responses may lead to better understanding of disease processes and suggest new therapeutic approaches. We introduce a method to detect such mediating functions using topological properties of protein-protein interaction networks. We introduce the concept of pathway centrality, a measure of communication between disease genes and differentially expressed genes. We find mediating pathways for three pulmonary diseases (asthma; bronchopulmonary dysplasia (BPD); and chronic obstructive pulmonary disease (COPD)) using pathway centrality. Mediating pathways shared by all three pulmonary disorders heavily favor inflammatory or immune responses and include specific pathways such as cytokine production, NF Kappa B, and JAK/STAT signaling. Disease-specific mediators, such as insulin signaling in BPD or homeostasis in COPD, are also highlighted. We support our findings, some of which suggest new treatment approaches, both with anecdotal evidence from the literature and via systematic evaluation using genetic interactions.