Statistical modeling of single-cell epitranscriptomics enabled trajectory and regulatory inference of RNA methylation

The scDART-seq data used in the study "Statistical Modeling of Single-Cell Epitranscriptomics Enabled Trajectory and Regulatory Inference of RNA Methylation" was obtained from both the SMART-seq2 and 10x Genomics platforms. The processed data have been deposited on Figshare: https://doi.org/10.6084/m9.figshare.25815862. This repository contains all relevant code organized into three main sections: simulation code, real data analysis for the SMART-seq2 platform, and real data analysis for the 10x Genomics platform. Simulation Code The simulation section consists of a comprehensive R Markdown file (simulation.Rmd) that contains all the code necessary for the simulation components. SMART-seq2 Data Analysis The analysis of the SMART-seq2 dataset is divided into three parts: preprocessing of the original raw data obtained from PRJNA750145, comparison of m6A site detection performance, and case studies. Preprocessing preprocess_data.txt: This file outlines the pipeline for extracting m6A signals from the scDART-seq dataset on the SMART-seq2 platform. preprocess.Rmd: This R Markdown file processes the raw data for subsequent analysis. Site Prediction and KS Test compare_random_2_4_8_16_without_TRESS.Rmd: This file compares various methods (excluding TRESS) for m6A detection using randomly selected control samples. compare_random_2_4_8_16_TRESS.Rmd: This file uses the TRESS method for m6A detection with randomly selected control samples. compare_sim_2_4_8_16_without_TRESS.Rmd: This file evaluates different methods (excluding TRESS) for m6A detection using control samples with similar expression profiles to the test cells. compare_sim_2_4_8_16_TRESS.Rmd: This file employs the TRESS method for m6A detection with similar expression profile control samples. ks_test_TRESS.Rmd: This file calculates the KS statistics for the TRESS method. ks_test_without_TRESS.Rmd: This file calculates the KS statistics for alternative methods. plot_code.Rmd: This file visualizes the results. Case Studies Case Study 1: 4_sim_all_test_cell: This directory contains four R Markdown files that generate differential methylation results for all 991 test cells using DESeq2, edgeR, QNB, and SigRM methods. trajectory: This file conducts trajectory analysis based on the results from the above methods. iM6A: This file verifies m6A findings using iM6A data. Case Study 2: pre-process_data.Rmd: This file preprocesses data from the results of Case Study 1. processed_data_185.Rmd: This file processes the data for further analysis. cnmf_plot.Rmd, cnmf.py: This file implements methods and visualizations for the analysis. 10x Genomics Data Analysis The analysis of the 10x Genomics dataset is similarly divided into two parts: preprocessing of the original raw data obtained from PRJNA750145, and a comparison of m6A site detection performance. Preprocessing preprocess_data.txt: This file outlines the pipeline for extracting m6A signals from the scDART-seq dataset on the 10x Genomics platform. preprocess.Rmd: This R Markdown file processes the raw data for subsequent analysis. Site Prediction and KS Test compare_rep.Rmd: This file compares the performance of different methods for m6A detection.