Understanding how genes are turned on and off in individual cells is essential for studying development and disease. Most single cell RNA sequencing methods focus on polyadenylated RNA, which mainly captures protein coding transcripts, but misses many important noncoding RNAs that play key regulatory roles.

In this work, researchers at Stanford University have developed a new approach called TotalX that expands what RNA sequencing can detect. This method allows scientists to capture both coding and noncoding RNA molecules in a single experiment using standard droplet-based platforms.

TotalX enables scalable detection of coding and noncoding RNAs, including miRNAs, in single cells

a, Schematic overview of the TotalX protocol. Total RNA is polyadenylated and reverse transcribed using a custom template-switching oligo (dUTSO). After reverse transcription, the TSO is digested with UDG and rRNA is depleted at the pre-amplified (Pre-amp) cDNA level using Cas9-based DASH. Long (>400 bp) and short (<400 bp) fragments are indexed separately, with optional inclusion of a gel-purified miRNA fraction (~18–50 bp) and pooled for sequencing. b, Gene detection efficiency across technologies. Comparison of the average number of genes per cell as a function of UMIs for TotalX (green), VASA-seq (orange) and 10x Genomics 3′ chemistry (blue), across binned depths. c,Unique genes detected per cell after UMI downsampling. Gene detection following normalization to 20,000 UMIs per cell. TotalX yields high gene complexity similar to VASA-seq and higher than the standard 10x Genomics Chromium 3′ workflow. The central line indicates the median, the box denotes the interquartile range (25th–75th percentiles) and the whiskers extend to the minimum and maximum values. The width of each violin represents the kernel density of the data. Sample sizes are indicated above each plot. d, Total number of unique genes detected per RNA biotype. Radial plots show numbers of unique genes detected in a representative experiment for each method, broken down by RNA biotype: protein-coding RNA, lncRNA, miscellaneous RNA (miscRNA), miRNA, snoRNA, snRNA, tRNA and histone RNA. Ratios represent the proportion of detected genes relative to the total number of annotated genes within each biotype. Only genes detected in 10 or more cells were counted. e, Improved detection of miRNAs using mixed library input. The scatterplot shows the average counts per million (CPM) per cell of TotalX alone (x-axis) versus TotalX with an added miRNA fraction (TotalX miRNA(+)) (y-axis) in HEK293T cells. Known HEK293T-specific miRNAs (red) reach expression levels comparable to those of low- and moderate-protein-coding genes (blue). Inset: proportion of reads mapping to the genome, indicating a trade-off with miRNA inclusion. f, Coverage profiles for selected miRNAs. Read depth plots for MIR17, MIR222 and MIR221 showing mature miRNA arms (gray regions).

By applying this approach to the developing human brain, the team created a detailed map of RNA activity across many different cell types. They found that different classes of RNA, including microRNAs and transfer RNAs, show distinct patterns depending on the cell type and developmental stage. This provides a more complete picture of how cells are regulated.

One important finding involved a microRNA called MIR137, which has been linked to neurological conditions such as schizophrenia. The researchers observed how its expression changes over time in specific neuron populations and how it interacts with its target genes, suggesting tight control of gene regulation during brain development.

The method was also tested in other systems, including blood cells and virus infected liver cells. In these cases, TotalX was able to detect RNA molecules that are typically missed, including viral RNA that lacks polyadenylation. This makes it a powerful tool for studying infections and immune responses.

By expanding RNA sequencing to include a broader range of RNA types, this approach provides deeper insight into how cells function and respond to their environment. It also opens new opportunities for building more comprehensive cellular atlases that reflect the full complexity of gene regulation.

Availability – Custom analysis scripts and instructions for Cell Ranger modifications are available via GitHub and Zenodo at https://doi.org/10.5281/zenodo.18177678