FORT LAUDERDALE, FLA. (GenomeWeb) – Researchers at the University of Michigan have started to use the Drop-Seq method for single cell transcriptomics in pilot studies to analyze large numbers of human kidney cells, presenting initial results of their efforts at the Association for Biomolecular Resource Facilities annual meeting here this week.
The Drop-Seq approach, a nanodroplet-based method developed by researchers at Harvard Medical School, the Broad Institute, and elsewhere that was published last year, allows scientists to barcode and analyze the messenger RNA of tens of thousands of single cells in parallel.
In his talk, Edgar Otto of the University of Michigan described how his group is beginning to use Drop-Seq to study kidney cells, in particular to determine what a healthy kidney cell transcriptome looks like and how it may change in disease.
Similar to the Broad group in its paper last year, their first pilot study was of a 50-50 mix of human and mouse cells to test the sensitivity and accuracy of the method. This, he told GenomeWeb in an interview, "worked nicely, but our interest of course is kidneys."
So far, his team has sequenced one library generated from a kidney sample and has additional libraries pending.
This first sample set was of healthy kidney tissue that had been obtained from a patient who had undergone a nephrectomy. While that surgery was to remove a tumor, a small amount of healthy tissue was also taken out. Otto then analyzed between 100,000 and 300,000 single cells from that sample using Drop-Seq, with the sequencing performed on the Illumina MiSeq.
In this pilot study, he said that about 5 percent of droplets had beads, and of those, 1 percent to 2 percent contained two beads.
During his talk, he said that even when he was unable to get a very high sequencing depth and could only detect the expression of a hundred genes, he could still cluster certain cell types together based on their expression.
"I think that the method is really promising for any tissue," Otto said, especially patient tissue where scientists want to see changes in expression levels for disease genes.
Indeed, Otto told GenomeWeb that he plans to conduct a study similar to what the Harvard-led team did in retinal cells. The Harvard team uncovered 39 different cell populations within the retina based on their transcriptional profiles, and Otto hopes to do the sample based on adult kidney cells transcriptomes.
Eventually, he could analyze the transcriptomes of kidney cells collected from patients with chronic kidney disease to isolate which cell types are responsible for the disease state or have changed transcripts in disease. Those transcripts, he added, could be biomarkers for disease as well as help identify biological pathways involved in the disease.
Otto noted that he has had some problems using frozen tissue. In particular, each time tissues are frozen and thawed, only a portion of cells survive the process.
While there have been improvements in that area, in some instances, only 20 percent of cells may survive to be analyzed. Then as Drop-Seq only tags a portion of those cells, that means an even smaller portion of cells are actually analyzed. "Drop-Seq isn't the best method for these precious samples," he said.
For such small or precious samples, he said that the inDrop approach developed at the same time at Harvard, and based on similar technology, might be a better approach, though he has yet to test it.
Still, the Drop-Seq approach is relatively inexpensive. Otto calculated that his setup costs were about $10,000, with another $10,000 in chemicals and consumables. He further estimated the per-cell transcript cost to be about 10 cents, though he noted that it depends largely on sequencing depth.