Digital Cell Sorter (DCS): a cell type identification, anomaly detection, and Hopfield landscapes toolkit for single-cell transcriptomics

Abstract

Motivation: Analysis of singe cell RNA sequencing (scRNA-seq) typically consists of different steps including quality control, batch correction, clustering, cell identification and characterization, and visualization. The amount of scRNA-seq data is growing extremely fast, and novel algorithmic approaches improving these steps are key to extract more biological information. Here, we introduce: (i) two methods for automatic cell type identification (i.e. without expert curator) based on a voting algorithm and a Hopfield classifier, (ii) a method for cell anomaly quantification based on isolation forest, and (iii) a tool for the visualization of cell phenotypic landscapes based on Hopfield energy-like functions. These new approaches are integrated in a software platform that includes many other state-of-the-art methodologies and provides a self-contained toolkit for scRNA-seq analysis. Results: We present a suite of software elements for the analysis of scRNA-seq data. This Python-based open source software, Digital Cell Sorter (DCS), consists in an extensive toolkit of methods for scRNA-seq analysis. We illustrate the capability of the software using data from large datasets of peripheral blood mononuclear cells (PBMC), as well as plasma cells of bone marrow samples from healthy donors and multiple myeloma patients. We test the novel algorithms by evaluating their ability to deconvolve cell mixtures and detect small numbers of anomalous cells in PBMC data. Availability: The DCS toolkit is available for download and installation through the Python Package Index (PyPI). The software can be deployed using the Python import function following installation. Source code is also available for download on Zenodo: doi.org/10.5281/zenodo.2533377.

Appyters: Turning Jupyter Notebooks into data-driven web apps

“… 64 Tempora is a pathway-based single-cell trajectory inference method that infers the developmental lineage of single cells based on pathway information. To predict cell types, Digital Cell Sorter 65 is implemented to categorize single cells into their hematopoietic lineage. In the future, we plan to add additional cell type prediction algorithms.…”

Naturally occurring combinations of receptors from single cell transcriptomics in endothelial cells

“…Two versions of the software implementation of the algorithms have been written independently by two of the authors (SD and AH), to guard against coding errors and unforeseen consequences of implementation choices. We have also used two different methods to identify endothelial cells in single cell RNAseq data: DCS, (20,21) which was previously developed by some of us, and Alona (19). Other validation strategies also represent replications in a broader sense.…”

“…This cell type annotation is incorporated in the PanglaoDB database. In addition to the Alona annotation, we processed each of the datasets in PanglaoDB using a platform for cell type annotation that we have recently developed (20,21), DCS 1.3.6.10.…”

“…We used the DCS pipeline (20,21) to process the single-cell RNA-seq PanglaoDB database (18). The pipeline was applied to each sample independently, without correcting for batch effects, to normalize the data and identify endothelial and non-endothelial cells.…”

Community assessment of methods to deconvolve cellular composition from bulk gene expression

“…S2 ; Table S13 ). mitten_TDC19 calculates a summarization score as the sum of the expression of selected markers, with the cell type-specific markers first nominated from expression profiles in purified bulk data or identified 34,35 from single-cell data expression profiles and then prioritized according to their correlation with that cell type's proportion over synthetic admixtures (Supplemental Methods; https://github.com/sdomanskyi/mitten_TDC19 ; Table S14 ). Finally, Biogem, based on a previously published method, 36 uses robust linear modeling to perform deconvolution and differential expression-based feature selection to define the purified expression profiles (Supplemental Methods; https://github.com/giannimonaco/DREAMChallenge_Deconvolution ; Fig.…”