Randall Lewis

Large-scale in silico modeling of metabolic interactions between cell types in the human brain

Metabolic interactions between multiple cell types are difficult to model using existing approaches. Here we present a workflow that integrates gene expression data, proteomics data and literature-based manual curation to model human metabolism within and between different types of cells. Transport reactions are used to account for the transfer of metabolites between models of different cell types via the interstitial fluid. We apply the method to create models of brain energy metabolism that recapitulate metabolic interactions between astrocytes and various neuron types relevant to Alzheimers disease. Analysis of the models identifies genes and pathways that may explain observed experimental phenomena, including the differential effects of the disease on cell types and regions of the brain. Constraint-based modeling can thus contribute to the study and analysis of multicellular metabolic processes in the human tissue microenvironment and provide detailed mechanistic insight into high-throughput data analysis.A workflow is presented that integrates gene expression data, proteomic data, and literature-based manual curation to construct multicellular, tissue-specific models of human brain energy metabolism that recapitulate metabolic interactions between astrocytes and various neuron types. Three analyses are applied for gene identification, analysis of omics data, and analysis of physiological states. First, we identify glutamate decarboxylase as a target that may contribute to cell-type and regional specificity in Alzheimer’s disease. Second, the decreased metabolic rate seen in affected brain regions in Alzheimer’s disease is consistent with a suppression of central metabolic gene expression in histopathologically normal neurons. Third, we identify pathways in cholinergic neurons that couple mitochondrial metabolism and cytosolic acetylcholine production, and subsequently find that cholinergic neurotransmission accounts for ∼3% of brain neurotransmission. Constraint-based modeling can thus contribute to the study and analysis of multicellular metabolic processes in human tissues, and provide detailed mechanistic insight into high-throughput data analysis.

Display advertising impact: search lift and social influence

We study the impact of display advertising on user search behavior using a field experiment. In such an experiment, the treatment group users are exposed to some display advertising campaign, while the control group users are not. During the campaign and the post-campaign period we monitor the user search queries and we label them as relevant or irrelevant to the campaign using techniques that leverage the bipartite query-URL click graph. Our results indicate that users who are exposed to the advertising campaign submit 5% to 25% more queries that are relevant to it compared to the unexposed users. Using the social graph of the experiment users, we also explore how users are affected by their friends who are exposed to ads. Our results indicate that a user with exposed friends is more likely to submit queries relevant to the campaign, as compared to a user without exposed friends. The result is surprising given that the display advertising campaign that we study does not include any incentive for social action, e.g., discount for recommending friends.