The majority of the organisms in the lungs are expected to have no effect, and severe asthmatics are expected to have more pathogenic and fewer beneficial microbes. A simplified model assigns microbes to one of three categories: pathogenic organisms that cause inflammation, beneficial organisms that reduce inflammation, and commensal organisms that have no effect on inflammation. Some fungi have well-defined effects on asthma, but the role of many microbes remains uncertain.

Some of these subgroups respond differently to environmental and microbial triggers, such as fungal spores. introduced transcriptional endotypes of asthma, and the Severe Asthma Research Program defined five subtypes of asthma. Much work has focused on identifying subgroups of the disease and how each subgroup responds to treatment. Our approach is useful when the heterogeneity comes from the sample type (e.g., sputum) and especially when the samples derive from a heterogeneous population of individuals, such as patients with asthma.įull size image Interactions between the host and microbes in the lungĪsthma is a disease of the airway that can present with diverse clinical phenotypes. These techniques can produce a low-dimensional representation of common groups of genes, microbes, or other features that tend to increase or decrease in abundance together. Here, we present a strategy for dealing with the complexity that uses a number of supervised and unsupervised techniques such as cell-type signatures and latent Dirichlet allocation (LDA). Moreover, generating RNA-seq data from the sputum can be technically easier than other methods to study microbe and immune cell interactions, such as by culturing sputum microbes or by flow sorting sputum immune cells. But within this complexity is an opportunity-the interactions between immune cells and microbes may provide clinically meaningful information from a non-invasive sample. By contrast, sputum may be less consistently collected, its cell types are less well-defined, and it may contain RNA from microbes and other organisms that act as cryptic indicators of the environment. In the blood, the vast majority of RNA-seq reads align to the human genome, and the goal is often to relate the expression of the genes to a phenotype. This can be appreciated by comparing sputum RNA-seq to a more traditional experiment, e.g., blood RNA-seq, where the sample can be collected consistently and that contains relatively well-defined cell types (Fig.

A key problem tackled in this paper is dealing with the heterogeneity and noise in RNA-seq data in samples such as sputum. Depending on the sample type, these data can be very heterogeneous. RNA sequencing (RNA-seq) has become a standard method of analyzing complex communities.

Linking high-dimensional, heterogeneous datasets

Quite revealing 1.8e serial