Flufftail: Fuzzy Learning and Uncertainty Framework For Transcriptional Trajectories And Interaction Logic.
Thanks to recent, rapid technological advances, single-cell assays are now state-of-the-art for a wide range of biomedical projects; RNA-focused assays represent a cost-effective and quick proxy for both DNA and protein quantifications, in addition to intrinsic expression signal. Clustering is pivotal in scRNA-seq for partitioning cells into distinct groups (putative cell types), setting the stage for analyses such as trajectory inference or cell-cell interactions. Recently, community detection algorithms emerged as superior/efficient for identifying transcriptomic-driven subpopulations.
Despite their enhanced performance, all cutting-edge community detection techniques are inherently stochastic and crisp; iterative runs result in variable partitions (and interpretations) when applied to the same input, and with the same hyper-parameters. This variability highlights the need of replacing crisp assignation with a probabilistic/fuzzy approach i.e., fuzzy clustering, where cells can concurrently be associated with multiple clusters.
Inspired by findings from Gribben et al. on biphenotypic cells in metabolic dysfunction-associated diseases, we introduce Flufftail, a comprehensive R package designed for single-cell datasets (agnostic to modalities), summarising signal through the lens of fuzzy clustering. Flufftail exploits the variability of standard clustering approaches by proposing a fuzzy community-detection clustering coupled with the characterization of fuzzy entries (cells/genes). The assessment of membership degrees, characterisation of hard clusters, and evaluation of co-clustering behaviour are summarised in interactive plots, facilitating the information transfer between wet- and dry-lab scientists. Additionally, we developed a new methodology for identifying key genes (major regulatory hubs) that drive biological transitions through fuzzy gene module clustering. Furthermore, Flufftail presents a new approach for characterising gene regulatory network (GRN) dynamics and the evolution of regulatory interactions across the pseudotime ordering of cells.
This marks the first exploration/exploitation of GRN dynamics across different points (or bins) of pseudotime in single-cell data, advancing our understanding of the mechanisms controlling cellular plasticity and transdifferentiation.
- Perform fuzzy clustering and examine results through:
- UMAP plots
- Hard clusters
- Membership degrees
- Consensus matrices
- Explore fuzziness metrics to:
- Identify fuzzy cells
- Highlight areas of fuzziness
- Conduct pseudotime analysis to:
- Trace cell development
- Identify transitioning clusters through the decomposed pseudotime plot
- Identify differentially expressed genes across the pseudotime area of transition cells between two clusters. If using the Shiny app can also see enriched pathways for the dataset.
- Perform fuzzy gene module clustering on the DE genes of the transition area to identify potential major regulatory hubs genes driving the dynamics in the transition
- Infer gene regulatory networks to observe dynamic changes across different points of the process for the hub genes (or specific genes within a selected pathway).
A main feature of Flufftail is its interactive analysis capability via an R Shiny interface. This extends its utility beyond bioinformaticians to include wet-lab researchers, enabling easier and faster exploration of data.
The Shiny application interface provides access to Flufftail's main functionalities. Each tab in the interface corresponds to different analysis capabilities.
A video demonstrating the analysis is available here: https://www.youtube.com/watch?v=C0JJMdQvRew&ab_channel=USN2292_cambridge.
For users who prefer to perform analyses through R scripts, Flufftail also provides a set of functions that mirror the capabilities of the Shiny interface. Please refer to the function documentation for more details.
Also the vignette "Flufftail Analysis on the Gribben et. al dataset" provides a detailed step-by-step analysis using the package.
The package will be later available through CRAN. To install it now you can run the following:
devtools::document()
devtools::build()
devtools::install()