Fig. 1: Systematic screening of small molecules for EF fate reprogramming using scRCF.
From: Chemical reprogramming of fibroblasts into retinal pigment epithelium cells for vision restoration
a The workflow of scRCF. The input of scRCF comprises scRNA-seq data from both initial and target cell types, in addition to three pre-established databases: (1) A small molecule perturbation database, integrating data from GEO and LINCS L1000, retaining only TFs identified in AnimalTFDB3.0; (2) The small-molecule target database and classification information, incorporating information from STITCH, Drug Repurposing Hub, and MedChemExpress; (3) A signaling network database, derived from Reactome and OmniPath. scRCF contains three major steps: (1) Identify candidate signaling proteins and perform pre-screening using the signaling network; (2) Calculate efficiency scores for small molecules based on the community partitioning results and output the highest-scoring compounds from each category as candidate small molecules; (3) Further screen the candidate small molecules using DRUG-seq2. b UMAP visualization of integrated scRNA-seq data from four MEF datasets and one primary EF (pEF) dataset, used to predict small molecules that facilitate reprogramming between the two cell types. c Volcano plot illustrating the differentially expressed TFs (DETFs) identified from scRNA-seq data of MEFs and EFs, analyzed using Seurat. DETFs were defined by p value < 0.05 and log2 fold change >1. d Candidate small molecules identified by preliminary screening with scRCF for reprogramming MEFs into EF cells. e Bar plot showing the Z-scores for cells treated with LAC + 1 small molecule combinations, based on a gene set of neuroectoderm- and EF-related genes. The Z-scores indicate relative gene expression changes for each drug, with higher values reflecting a stronger effect on the gene set. Bars are ordered from the highest to lowest average Z-score, and the dashed line represents the baseline score of zero. f Heatmap showing the gene expression profiles of neuroectoderm- and EF-related genes in cells after treatment with LAC + 1 small molecule combinations. The color scale represents log2-transformed expression values, with red indicating high expression, blue indicating low expression, and white indicating intermediate levels.
