scrunchy provides analysis tools for the single-cell reconstruction of functional heterogeneity.

New methods to study heterogeneity at cellular resolution in complex tissues are rapidly transforming human biology. These methods measure differences in gene expression, chromatin accessibility, and protein levels across thousands of cells to understand developmental trajectories of tissues, tumors, and whole organisms. However, their reliance on measurements of steady-state abundance of DNA, RNA, and protein limits our ability to extract dynamic information from single cells.

To propel the study of heterogeneity among single cells, we are developing functional assays as a new modality for single-cell experiments. Instead of measuring the molecular abundance of DNA, RNA, or protein in single cells and predicting functional states, our key innovation is to directly measure enzymatic activities in single cells by analyzing the conversion of substrates to products by single-cell extracts in a high-throughput DNA sequencing experiment.

Functional heterogeneity of DNA repair in immune cells

Our first functional method simultaneously measures the activity of DNA repair enzymes and the abundance of mRNAs from thousands of single cells. We measure DNA repair activities by encapsulating synthetic DNA oligonucleotides with defined lesions with single cells. Cellular DNA repair enzymes recognize and catalyze incision of these substrates, which we subsequently capture in a modified library construction protocol.

An example data set in scrunchy contains a subset of data from an experiment in which we simultaneously measure mRNA expression and specific types of DNA repair activities in thousands of single cells (human peripheral blood mononuclear cells).

The following plot shows cells classified by mRNA expression for each cell type embedded in a two-dimensional UMAP projection. This data set is too small to robustly classify cell types, but the full data set (containing ~4,000 cells) indicates that clusters 1, 2, 3, and 5 are T-cell subtypes.

In addition to mRNA expression, we measured the activity of DNA repair factors in each of these cells. The following plots show DNA repair activities measured for each cluster above. The repair substrates included an unmodified DNA substrate (A), an A:U base-pair (B), a ribonucleotide (rG-D, C), and an abasic site (D). Activity is measured as a normalized count of incisions observed at expected repair positions.

These data show that whereas little repair activity is associated with the unmodified DNA (A), uracil base excision (B), ribonucleotide excision (C), and abasic site processing (D) activities can be measured in single cells. Moreover, differences in activity between clusters may reflect intrinsic differences in the levels of these DNA repair pathways.


scrunchy is under active development. You can install the R package from github: