Mapping a New Course to Understand Lung Biology Mechanisms: LungMAP.net

The lung is a highly complex human organ, harboring over 40 different cell types with specialized functions to facilitate gas exchange and host defense (1). Single-cell omics technologies, such as single-cell RNA sequencing (scRNA-seq), have been leveraged to create multiple atlas-level single-cell datasets of the lung (2). The functions of individual cells have been characterized using image resolution technologies, such as immunofluorescence confocal microscopy, to elucidate the spatial distribution of cells, proteins, mRNA, and metabolites. Although omics data promise additional critical insights into the cellular biology of the lung, integrating omics data with existing single-cell atlases and image resolution technologies has posed a challenge for many mainstream biologists. Web servers and computational platforms have been designed for the general analysis of lung omics data, with minimal prior computational knowledge to enhance the accessibility of omics data analysis for mainstream biologists (3, 4). However, with the growing number of lung and airway datasets, there is an increasing need for an integrative, user-friendly online tool to revolutionize the identification of disease mechanisms. This will aid the design of novel diagnostic and personalized therapeutic regimens for complex lung diseases such as pulmonary fibrosis, bronchopulmonary dysplasia, chronic obstructive pulmonary disease, asthma, and others.

In this issue of the Journal, Gaddis and colleagues (pp. 129–139) propose LungMAP.net as a new gateway portal for lung research that integrates computational tools and multi-omics data resources (Figure 1) (5). This comprehensive website links eight independent lung research portals with omics and image data, adopts a core set of community schemas to interconnect metadata and file formats for each portal, and provides applications to map the user’s data to reference panels. Considering the complexity of LungMAP.net, the authors provide 10 tutorials, including over 3 hours of video tutorials that cater to users with different backgrounds and varying degrees of expertise. This effectively democratizes access to advanced lung biology research.

Figure 1.

Overview of LungMAP.net. The LungMAP.net portal mainly includes two sections: (Left) data resource and (Right) visualization and analysis. In the data resource section, LungMAP aggregates a comprehensive collection of bulk and single-cell multi-omic data, encompassing transcriptomics, proteomics, epigenomics, metabolomics, and lipidomics from human, mouse, and rhesus specimens together with extensive imaging datasets. In the Visualization and Analysis section, CellCards offers detailed biological insights into various curated lung cell types and a cell type annotation web tool for human and mouse through Azimuth. The portal also facilitates a wide range of data visualization methods through ShinyCell, including Uniform Manifest Approximation and Projection (or UMAP), boxplots, violin plots, bubble plots, and heatmaps. Additionally, it integrates with external portals such as Lung Gene Expression Analysis (or LGEA), ToppCell, and Terra, further enhancing its utility for advanced lung single-cell genomics research. The figure was created by BioRender.com. OMERO = open microscopy environment; RNA-seq = RNA sequencing; TB = terabyte.

The main advantages of LungMAP.net include the following:

• 1.Interactive analysis. LungMAP.net provides interactive operations for lung data analysis and visualization. CellRef, a R Shiny application (Azimuth), allows users to map scRNA-seq data to references provided from CellCards (6). It integrates ShinyCell to visualize the clustering results and gene expression using boxplots, violin plots, heatmaps, and scatterplots (7). It also integrates ToppCell to curate gene signatures of cells on the basis of their lineages, cell classes, subclasses, clinical information, and other features (8). Users can reprocess the data from the LungMAP Data Browser through Terra using open-access single-cell analysis workflows in the cloud.
• 2.Comprehensive integration. LungMAP.net collected: 1) omics array data from 206 specimens of different ages, including bulk and scRNA-seq, proteomics, lipidomics, DNA methylation, microRNA, and metabolomics measurements for diverse samples, including those matched for the same samples (multi-omics); 2) image data to understand the spatial and cellular localization of proteins, mRNA transcripts, and metabolites in the lung; 3) CellCards that integrate functional and emerging single-cell genomic datasets for accurate and evolvable annotation of important cell lineages within the respiratory system (6); and 4) Lung Gene Expression Analysis (LGEA) that explores multi-omics and imaging data through an interactive web interface to connect lung structure and histology to cell type–specific gene expression (9). In addition, LungMAP.net uses a standardized and iterative curation metadata approach, supporting new technologies and sample types.

This harmonized central resource supports hypothesis-driven research questions and broad exploratory investigations. The authors define how users can query specific genes, proteins, cell types, datasets, diseases, or developmental stages to discover affected molecular and cellular programs for subsequent validation in independent cohorts. LungMAP.net breaks down silos and prevents researchers from reinventing the wheel by enhancing and refining data and analysis sharing, thereby reducing unnecessary redundancy and increasing opportunity for novel exploration and discovery. Integrative resources such as LungMAP.net thus have the potential to accelerate scientific discoveries that will ultimately translate into improved care for patients with pulmonary disease.

Although LungMAP.net integrates comprehensive multi-omics data, the ecosystem has two potential directions for future improvement. First, LungMAP.net could collect spatial transcriptomics data for lung diseases, which encompasses a suite of innovative technologies that enable positional profiling of gene expression in the lung (10, 11). Further spatial transcriptomics data analysis tasks, such as spatial domain detection and spatial variable gene detection, can be facilitated through the Terra cloud environment (12–14). Second, LungMAP.net could develop an additional plug-in or webpage to integrate multi-omics data, such as bulk and scRNA-seq with newly developed analytic tools in the field (15).

LungMAP.net is a comprehensive lung ecosystem with multi-omics, multitechnology, and multispecies information. It also utilizes a cloud computing platform to make integrative analysis available to physicians and biologists who lack significant prior experience with these methodologies. Aggregating data from various research networks and promoting broader scientific community participation and collaboration, LungMAP.net paves the way for enhanced molecular biological education and research for the lung.