They isolated the cells using an inverted microscope and microcapillary pipetting followed by whole-genome amplification (WGA) based on multiple-displacement amplification with the φ29 enzyme, which allowed linear amplification of the complete genome. WGA DNA was then analyzed by massively parallel sequencing with Illumina
instruments. The authors developed and validated the method with two single cells from a lymphoblastoid cell line isolated from the individual whose DNA was used for the first Asian diploid genome sequenced. This work showed that whole-genome sequencing of WGA DNA has acceptable error rates, and it provided a baseline for subsequent studies of single cells from cancer samples using only exome sequencing.
Next, Hou et al.2 sequenced the exomes of 58 single cancer cells from an individual with essential thrombocythemia, a myeloproliferative neoplasm, to a mean depth of 30×. Using mutation-calling algorithms, which were experimentally validated both by comparing all single-cell somatic variants to those identified in whole essential-thrombocythemia tissue sequencing and by PCR-Sanger sequencing of 30 randomly selected mutations, the authors identified 712 somatic variants, of which 171 coding variants were further assessed. Seventy-eight of these 171 coding somatic mutations (in a total of 71 genes) were nonsynonymous. A population-genetics analysis of these data revealed the monoclonal origin of essential-thrombocythemia cells in this patient. Furthermore, a driver-gene prediction model identified eight genes as the ones with the
highest likelihood of being involved in the
neoplastic initiation and/or progression of essential thrombocythemia.
In a second study from the same laboratory, Xu et al.3 applied the single-cell sequencing method to analyze the exomes of 25 single cells from a renal cell carcinoma—20 from the tumor and 5 from adjacent normal tissue.
The sequencing data revealed 260 coding somatic mutations. Principal component analysis showed that three of the single cancer
cells clustered tightly with the five normal single cells, which the authors took to indicate that these were actually healthy cells admixed within the tumor. The remaining 17 single cancer
cells had 229 somatic mutations, and these mutations were used to determine the allelic frequency at the 229 nucleotide positions. This analysis revealed two distinct peaks, one with a frequency range of 0–5% and the other with a frequency range of 15–20%. Thus, not only was there significant intratumoral heterogeneity, but also—notably—there were no dominant clones identified in the cancer tissue. In other words,