Sfaira accelerates data and model reuse in single cell genomics

Fischer, David; Dony, Leander; König, Martin; Moeed, Abdul; Zappia, Luke; Heumos, Lukas; Tritschler, Sophie; Holmberg, Olle; Aliee, Hananeh; Theis, Fabian J.

doi:10.1186/s13059-021-02452-6

Cited by 21 publications

(21 citation statements)

References 71 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The HLCA core is publicly available as a data portal and a model repository to explore, download, and use as a reference for new datasets. As the atlassing community has multiple outlets for newly generated data, we made the atlas available in Sfaira 70 , Zenodo 71 , Azimuth 14 , CellTypist 72 , and FASTGenomics 73 . Mapping new datasets to the HLCA core using scArches can be done via interactive portals, such as FASTGenomics and Azimuth, as well as locally using our Zenodo model 74 , which lends itself to integration into bioinformatics pipelines.…”

Section: Discussionmentioning

confidence: 99%

An integrated cell atlas of the human lung in health and disease

Sikkema

Strobl

Zappia

et al. 2022

Preprint

Self Cite

126

View full text Add to dashboard Cite

Organ- and body-scale cell atlases have the potential to transform our understanding of human biology. To capture the variability present in the population, these atlases must include diverse demographics such as age and ethnicity from both healthy and diseased individuals. The growth in both size and number of single-cell datasets, combined with recent advances in computational techniques, for the first time makes it possible to generate such comprehensive large-scale atlases through integration of multiple datasets. Here, we present the integrated Human Lung Cell Atlas (HLCA) combining 46 datasets of the human respiratory system into a single atlas spanning over 2.2 million cells from 444 individuals across health and disease. The HLCA contains a consensus re-annotation of published and newly generated datasets, resolving under- or misannotation of 59% of cells in the original datasets. The HLCA enables recovery of rare cell types, provides consensus marker genes for each cell type, and uncovers gene modules associated with demographic covariates and anatomical location within the respiratory system. To facilitate the use of the HLCA as a reference for single-cell lung research and allow rapid analysis of new data, we provide an interactive web portal to project datasets onto the HLCA. Finally, we demonstrate the value of the HLCA reference for interpreting disease-associated changes. Thus, the HLCA outlines a roadmap for the development and use of organ-scale cell atlases within the Human Cell Atlas.

show abstract

Section: Discussionmentioning

confidence: 99%

An integrated cell atlas of the human lung in health and disease

Sikkema

Strobl

Zappia

et al. 2022

Preprint

Self Cite

126

View full text Add to dashboard Cite

show abstract

“…To test this, we modified the scVI encoder to employ a non-amortized formulation (see Methods ), in which the parameters of the variational distribution are optimized for each cell individually. We trained expiMap, scVI, and non-amortized scVI to construct references with multiple atlases across five tissues obtained from Sfaira 53 . We observed that non-amortized versions of scVI consistently achieved superior or equal performance in data integration compared with expiMap, while remaining better than the default (amortized) scVI, which corroborated our hypothesis ( Fig.…”

Section: Resultsmentioning

confidence: 99%

“…We leverage datasets from five different tissues including PBMCs (n=161,764) [28], heart (n=18641) [30], lung (n=65,662) [31], colon (n=34,772) [32], and liver (n=113,063) [33]. All datasets, except heart, were obtained from the Sfaira database [34], which includes cell type labels. Heart was obtained from the scVI package.…”

Section: Expimap - Online Methodsmentioning

confidence: 99%

Biologically informed deep learning to infer gene program activity in single cells

Lotfollahi

Rybakov

Hrovatin

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

The increasing availability of large-scale single-cell datasets has enabled the detailed description of cell states across multiple biological conditions and perturbations. In parallel, recent advances in unsupervised machine learning, particularly in transfer learning, have enabled fast and scalable mapping of these new single-cell datasets onto reference atlases. The resulting large-scale machine learning models however often have millions of parameters, rendering interpretation of the newly mapped datasets challenging. Here, we propose expiMap, a deep learning model that enables interpretable reference mapping using biologically understandable entities, such as curated sets of genes and gene programs. The key concept is the substitution of the uninterpretable nodes in an autoencoder's bottleneck by labeled nodes mapping to interpretable lists of genes, such as gene ontologies, biological pathways, or curated gene sets, for which activities are learned as constraints during reconstruction. This is enabled by the incorporation of predefined gene programs into the reference model, and at the same time allowing the model to learn de novo new programs and refine existing programs durin reference mapping. We show that the model retains similar integration performance as existing methods while providing a biologically interpretable framework for understanding cellular behavior. We demonstrate the capabilities of expiMap by applying it to 15 datasets encompassing five different tissues and species. The interpretable nature of the mapping revealed unreported associations between interferon signaling via the RIG-I/MDA5 and GPCRs pathways, with differential behavior in CD8+ T cells and CD14+ monocytes in severe COVID-19, as well as the role of annexins in the cellular communications between lymphoid and myeloid compartments for explaining patient response to the applied drugs. Finally, expiMap enabled the direct comparison of a diverse set of pancreatic beta cells from multiple studies where we observed a strong, previously unreported correlation between the unfolded protein response and asparagine N-linked glycosylation. Altogether, expiMap enables the interpretable mapping of single cell transcriptome data sets across cohorts, disease states and other perturbations.

show abstract

“…Advancement towards ultra-high throughput single-cell RNA sequencing (scRNA-seq) platforms and the concomitant development of computational algorithms required to analyze the data, permits the generation of organism-wide transcriptomic maps, resolved both in time and space [1,2]. If a reference atlas is available, new datasets can be annotated automatically thus introducing fast, data-driven and consistent labeling of the cells [3][4][5]. Unfortunately, the skeleton is minimally represented in most of these atlases, often with insufficiently detailed annotation of the skeletal lineage.…”

Section: Introductionmentioning

confidence: 99%

An integrated single-cell atlas of the skeleton from development through adulthood

Herpelinck

Ory

Nasello

et al. 2022

Preprint

View full text Add to dashboard Cite

The recent growth of single-cell transcriptomics has turned single-cell RNA sequencing (scRNA-seq) into a near-routine experiment. Breakthroughs in improving scalability have led to the creation of organism-wide transcriptomic datasets, aiming to comprehensively profile the cell types and states within an organism throughout its lifecycle. To date however, the skeleton remains a majorly underrepresented organ system in organism-wide atlases. Considering how the skeleton not only serves as the central framework of the vertebrate body but is also the home of the hematopoietic niche and a central player in major metabolic and homeostatic processes, this presents a major deficit in current reference atlas projects. To address this issue, we integrated seven separate scRNA-seq datasets containing skeletal cells and their developmental precursors, generating an atlas of over 800,000 cells. This skeletal cell atlas describes cells across the mesenchymal lineage from the induction of the limb field to adult bone, encompassing 50 different cell states. In addition, the original datasets were reannotated, enabling the discovery of novel, highly specific marker genes, some of which we have validated in vivo by whole-mount in situ hybridization. Furthermore, expanding the repertoire of available time points and cell types within a single dataset allowed for more complete analyses of cell-cell communication or in silico perturbation studies. Taken together, we present a missing piece in the current atlas mapping efforts, which will be of value to researchers in the fields of skeletal biology, hematopoiesis, metabolism and regenerative medicine.

show abstract

Sfaira accelerates data and model reuse in single cell genomics

Cited by 21 publications

References 71 publications

An integrated cell atlas of the human lung in health and disease

An integrated cell atlas of the human lung in health and disease

Biologically informed deep learning to infer gene program activity in single cells

An integrated single-cell atlas of the skeleton from development through adulthood

Contact Info

Product

Resources

About