Velocity of single-cell epitranscriptome unveiled the transient dynamics, regulatory mechanisms and transcriptional impact of RNA modification

RNA modifications play critical roles in regulating gene expression, yet their dynamic behavior at single-cell resolution remains largely unexplored. To address this, we developed VeloRM, a computational framework that models the kinetics of RNA modifications—such as m⁶A methylation and A-to-I editing—by quantifying their gain and loss over pseudotime at single-nucleotide resolution. Applying VeloRM to scDART-seq data, we uncovered distinct regulatory programs: most m⁶A modifications occur prior to splicing, though a subset arises post-splicing; in contrast, A-to-I editing primarily takes place during or after splicing, following markedly different kinetics from m⁶A. VeloRM accurately reconstructs modification-associated cellular trajectories and identifies stage-specific modification patterns that influence RNA fate. Notably, m⁶A levels correlate with mRNA decay, with most regulatory effects observed in pre-mRNA, although some modification sites show a positive association with mature mRNA abundance. Together, these findings provide new insights into the dynamic regulation of the epitranscriptome at single-cell resolution and establish a powerful framework for studying RNA modification kinetics in development and disease.

1. Code

   The repository includes R Markdown scripts for analyzing both scDART-seq data and RNA-Seq A-to-I editing data.

   scDART-seq Data Analysis

   The analysis is divided into three main stages: (1) raw data preprocessing, (2) data processing, and (3) downstream analysis.

   1. Raw Data (origin)

   • SNP_all: SNP information for candidate m6A sites.
   • frequency_pre_YTH: Methylated and unmethylated read counts in pre-mRNA of test (YTH-WT) cells.
   • frequency_pre_YTHmut: Methylated and unmethylated read counts in pre-mRNA of control (YTH-mutant) cells.
   • frequency_spliced_YTH: Methylated and unmethylated read counts in mature mRNA of test cells.
   • frequency_spliced_YTHmut: Methylated and unmethylated read counts in mature mRNA of control cells.
   • frequency_isoform_YTH: Methylated and unmethylated read counts in ambiguous (isoform-level) mRNA of test cells.
   • frequency_isoform_YTHmut: Methylated and unmethylated read counts in ambiguous mRNA of control cells.
   • label_SigRM: Cell cycle stage labels from Statistical modeling of single-cell epitranscriptomics enabled trajectory and regulatory inference of RNA methylation (Cell Genomics).
   • expression_TPM: Gene expression data (TPM) from the same study.
   • gene_informations: Gene annotation data from the same study.

   2. Preprocessing

   • SNP_37199: Subset of SNPs used for downstream m6A analysis.
   • test_list: Processed read counts for test cells, including:
   • • spliced_meth_test, spliced_unmeth_test
     • pre_meth_test, pre_unmeth_test
     • isoform_meth_test, isoform_unmeth_test
   • control_list: Processed read counts for control cells, similarly structured.
   • res_preprocess: Preprocessed input formatted for VeloRM.
   • res_DESeq2: Differential expression analysis results using DESeq2.
   • SigRMtest_res_unspliced: SigRM test results on unspliced reads (test vs. control).
   • SigRMtest_res_spliced: SigRM test results on spliced reads (test vs. control).

   3. Analysis

   • splicing_junction_analysis: Data frame with distances between SNPs and the nearest splicing junctions.
   • res_m6A_velocity: m6A velocity results computed by VeloRM.
   • res_expression_velocity: Site-level expression velocity estimated by VeloRM.
   • res_transcriptional_impact_analysis_spearman: Transcriptional impact results (Spearman correlation-based) from VeloRM.
   • cell_cycle_sites: Cell cycle-related methylation site analysis performed with VeloRM.

   RNA-Seq A-to-I Data Analysis

   1. Raw Data (origin)

   • SNP_all: SNP information of candidate A-to-I editing sites.
   • pre_A-to-I: Read counts (methylated and unmethylated) in pre-mRNA of test cells.
   • spliced_A_to_I: Read counts in mature mRNA of test cells.
   • isoform_A_to_I: Read counts in ambiguous mRNA of test cells.

   2. Preprocessing

   • SNP_1159: Filtered SNPs for downstream A-to-I analysis.
   • test_list: Read counts for test cells, including:
   • • spliced_meth_test, spliced_unmeth_test
     • pre_meth_test, pre_unmeth_test
     • isoform_meth_test, isoform_unmeth_test
   • control_list: NULL (no control group in this dataset).
   • res_preprocess: Preprocessed data formatted for VeloRM.
   • res_DESeq2: Differential analysis results via DESeq2.

   3. Analysis

2. • splicing_junction_analysis: Distances between candidate sites and nearest splice junctions.
   • cell_cycle_sites: Cell cycle-related A-to-I site analysis via VeloRM.