Diagnostic Evidence GAuge of Single cells (DEGAS): A flexible deep-transfer learning framework for prioritizing cells in relation to disease

Abstract

With the rapid advance of single cell sequencing techniques, single cell molecular data are quickly accumulated. However, there lacks a sound approach to properly integrate single cell data with the existing large amount of patient-level disease data. To address such need, we proposed DEGAS (Diagnostic Evidence GAuge of Single cells), a novel deep transfer-learning framework which allows for cellular and clinical information, including cell types, disease risk, and patient subtypes, to be cross-mapped between single cell and patient data, provided they share at least one common type of molecular data. We call such transferrable information "impressions", which are generated by the deep learning models learned in the DEGAS framework. Using eight datasets from a wide range of diseases including Glioblastoma Multiforme (GBM), Alzheimer's Disease (AD), and Multiple Myeloma (MM), we demonstrate the feasibility and broad applications of DEGAS in cross-mapping clinical and cellular information across disparate single cell and patient level transcriptomic datasets. Specifically, we correctly mapped clinically known GBM patient subtypes onto single cell data. We also identified previously known neuron loss from AD brains, then mapped the "impression" of AD risk to single cell data. Furthermore, we discovered novel differences in excitatory and inhibitory neuron loss in AD data. From the exploratory MM data, we identified differences in the malignancy of different CD138+ cellular subtypes based on "impressions" of relapse information transferred from MM patients. Through this work, we demonstrated that DEGAS is a powerful framework to cross-infer cellular and patient-level characteristics, which not only unites single cell and patient level transcriptomic data by identifying their latent links using the deep learning approach, but can also prioritize both patient subtypes and cellular subtypes for precision medicine.

Insights into high-risk multiple myeloma from an analysis of the role of PHF19 in cancer

“…These data suggest that PHF19 overexpression in PC may re-induce the expression of genes involved in proliferation and cell cycle, a situation that would be biologically relevant to the malignant behavior of a cell type otherwise destined for quiescence or apoptosis when the initial immune response is no longer required. Furthermore, single cell RNA-seq analysis identified that only a subpopulation of MM malignant cells overexpress PHF19 [115]. Interestingly, this PHF19high myeloma cell subpopulation seems to be the cell cycling fraction as they also exhibited high expression of genes Fig.…”

“…This observation suggests that only a subpopulation of MM tumor cells express high levels of PHF19 bringing the association between PHF19 and MM aggressiveness to a cellular level. The presence of the PHF19high myeloma cell subtype-2 is showed to be associated with myeloma progression using a new deep learning tools called DEGAS [115]. Interestingly, this PHF19 high myeloma cell subtype also exhibited high expression of genes involved in cell cycle including HELLS, EZH2, TYMS, ZWINT, and MKI67 [115].…”

“…The presence of the PHF19high myeloma cell subtype-2 is showed to be associated with myeloma progression using a new deep learning tools called DEGAS [115]. Interestingly, this PHF19 high myeloma cell subtype also exhibited high expression of genes involved in cell cycle including HELLS, EZH2, TYMS, ZWINT, and MKI67 [115].…”

“…Along with other genes in this program (e.g., CKS1B, PCNA, E2F1, FOXM1), PHF19 expression was associated with cell proliferation and was among the best genes predictors of high-risk status [114]. In a recent preprint study analyzing single cell RNA-seq of FACS-sorted CD138+ MM cells using Seurat, showed that MM cells cluster into five major subtypes (clusters) with the cell subtype-2 expressing high level of PHF19 compared to the others subtypes [115]. This observation suggests that only a subpopulation of MM tumor cells express high levels of PHF19 bringing the association between PHF19 and MM aggressiveness to a cellular level.…”

“…ee/ gprofi ler/ gost). D Venn diagram showing the overlap between the 294 genes identified as the PHF19 transcriptional signature of MM and the 209 genes that positively correlate with expression of PHF19 involved in cell cycling including HELLS, EZH2, TYMS, ZWINT, and MKI67 [115].…”