Machine learning based classification of cells into chronological stages using single-cell transcriptomics

Abstract

Age-associated deterioration of cellular physiology leads to pathological conditions. The ability to detect premature aging could provide a window for preventive therapies against age-related diseases. However, the techniques for determining cellular age are limited, as they rely on a limited set of histological markers and lack predictive power. Here, we implement GERAS (GEnetic Reference for Age of Single-cell), a machine learning based framework capable of assigning individual cells to chronological stages based on their transcriptomes. GERAS displays greater than 90% accuracy in classifying the chronological stage of zebrafish and human pancreatic cells. The framework demonstrates robustness against biological and technical noise, as evaluated by its performance on independent samplings of single-cells. Additionally, GERAS determines the impact of differences in calorie intake and BMI on the aging of zebrafish and human pancreatic cells, respectively. We further harness the classification ability of GERAS to identify molecular factors that are potentially associated with the aging of beta-cells. We show that one of these factors, junba, is necessary to maintain the proliferative state of juvenile beta-cells. Our results showcase the applicability of a machine learning framework to classify the chronological stage of heterogeneous cell populations, while enabling detection of candidate genes associated with aging.

A single-cell atlas of de novo β-cell regeneration reveals the contribution of hybrid β/δ-cells to diabetes recovery in zebrafish

“…Collection of β-, δ1-, and hybrid cells for transcriptional profiling was performed using FACS of single-cell suspension of dissected islets from Tg(ins:BB1.0L); Tg(sst1.1:EGFP-Ras) . For single-cell profiling using Smart-Seq., cells were collected in 96-well plate using index sorting and processed according to the previously published pipeline (Singh et al, 2018). Analysis was performed using Seurat (Satija et al, 2015), with cells classified using their position on the FACS plot (δ1-cells: Green+, Red-; hybrid cells: Green+, Red+; and β-cells: Green-, Red+).…”

Single-cell transcriptome analysis reveals cell-cell communication and thyrocyte diversity in the zebrafish thyroid gland

“…Single cell suspension of zebrafish thyroid gland was performed by adapting the cell dissociation protocol outlined in Singh et al, Scientific Reports, 2018 (46). In brief, the thyroid gland was collected and dissociated into single cells by incubation in TrypLE (ThermoFisher, 12563029) with 0.1% Pluronic F-68 (ThermoFisher, 24040032) at 37 °C in a benchtop shaker set at 450 rpm for 45 min.…”

Machine learning in clinical decision making

“…29,30 In addition, methods exist to infer cellular states, e.g., aging, or responding versus non-responding cells upon treatment. 31,32 Human peripheral blood holds great promise as a potential ''liquid biopsy.'' The serum metabolome determines various diseases and was recently predicted with an ML model based on host genetics, gut microbiome, clinical parameters, diet, lifestyle, and anthropometric measurements.…”