Skip to main content
Premium Trial:

Request an Annual Quote

ABRF Speakers Describe Combining Sequencing With Other Approaches to Get More Data From Single Cells

MYRTLE BEACH, SC (GenomeWeb) – Researchers are combining 'omic and other experimental approaches to squeeze the most data out of their single cell samples.

In a session on the analysis of single cells at the annual Association of Biomolecular Resource Facilities meeting in Myrtle Beach, South Carolina on Tuesday, researchers described approaches they've recently developed that join multiple modalities together. One approach combined genomics and transcriptomics, while another joined transcriptomics and proteomics, and yet other married transcriptomics, cell morphology, and electrophysiology to piece together a more in-depth view of cells.

"Multi-ome assays are developing very, very fast," the Earlham Institute's Iain Macaulay said during his talk.

For instance, he and his colleagues have combined genomics and transcriptomics together in an approach they have dubbed G&T-seq. This, he said, was driven in part by cancer researchers who were interested in observing the effect of CNVs on expression by linking the two dataset types.

Their G&T-seq approach involves isolating cells, such as through flow cytometry, then lysing them, and attaching a bead to the mRNA to separate it out from the genomic DNA. The mRNA and DNA are then amplified separately and sequenced.

He and his colleagues applied their approach to breast cancer epithelial and lymphoblastoid B cell lines that were derived from the same patient. As they reported recently in Nature Methods, the researchers were able to detect a novel fusion that connects MTAP on chromosome 9 with PCDH on chromosome 4 in about 21 percent of the HCC38 cells within their RNA sequencing data. When they then performed deep, long-read sequencing of four HCC38 cells, they identified the rearrangement in three of those cells.

Macaulay added in his talk that he and his colleagues have also been able to use their approach to examine how regulatory changes affect transcription.

Similarly, the New York Genome Center's Marlon Stoeckius described in his talk an approach he and his colleagues developed called cellular indexing of transcriptomes and epitopes by sequencing, or CITE-seq, which merges transcriptomics and protein analysis. As they reported last summer in Nature Methods, CITE-seq involves encapsulating single cells, lysing them, using antibody-derived oligos to tag proteins before barcoding both them and mRNA, and separating the two for sequencing.

When he and his colleagues performed both CITE-seq and flow cytometry on peripheral blood mononuclear cells, they found that they similarly identified subsets of immune cells with the same relative distributions.

In his talk, he also described an extension of this approach, called Cell Hashing, that enables single-cell multiplexing. In this approach, oligo-tagged antibodies against cell surface proteins are used to label cells from different samples before pooling them for transcriptomic analysis. They also recounted this method in a preprint on BioRxiv.

Lastly, some researchers are folding in other experimental approaches with single-cell sequencing. The University of California, San Francisco's Cathryn Cadwell described during the session how she and her colleagues combined patch clamping, a mainstay of neuroscience research, with morphology analysis and sequencing to profile single neuronal cells.

In patch clamping, a pipette containing a solution similar to the cytoplasm is inserted into the cell membrane, and a wire in the pipette is used to record the electrophysiology of the cell. To add sequencing into the process, Cadwell said that when the pipette is removed, it sucks the cell contents out for subsequent analysis. At the same time, a dye spreads through the cell to uncover its morphology.

However, Cadwell noted that the process is laborious and not easy to automate. "It's a lot of effort to get a sample," she said, adding that they are able to get all three types of data — electrophysiology, morphology, and transcription — about 36 percent of the time.

Still, as she and her colleagues have reported in Nature Biotechnology, the cells' gene expression profiles can be used to predict their physiology. In particular, they reported generating both electrophysiological and molecular profiles for 58 neocortical cells. When they clustered the cells based on their transcriptional profiles, it was highly similar to clusters generated based on their electrophysiological profiles.

"There's been an explosion in many fields … of single-cell sequencing," she added.