Comparative study of whole genome amplification and next generation sequencing performance of single cancer cells

Abstract

Background: Whole genome amplification (WGA) is required for single cell genotyping. Effectiveness of currently available WGA technologies in combination with next generation sequencing (NGS) and material preservation is still elusive.

Results: In respect to the accuracy of SNP/mutation, indel, and copy number aberrations (CNA) calling, the HiSeq2000 platform outperformed IonProton in all aspects. Furthermore, more accurate SNP/mutation and indel calling was demonstrated using single tumor cells obtained from EDTA-collected blood in respect to CellSave-preserved blood, whereas CNA analysis in our study was not detectably affected by fixation. Although MDA-based WGA yielded the highest DNA amount, DNA quality was not adequate for downstream analysis. PCR-based WGA demonstrates superiority over MDA-PCR combining technique for SNP and indel analysis in single cells. However, SNP calling performance of MDA-PCR WGA improves with increasing amount of input DNA, whereas CNA analysis does not. The performance of PCR-based WGA did not significantly improve with increase of input material. CNA profiles of single cells, amplified with MDA-PCR technique and sequenced on both HiSeq2000 and IonProton platforms, resembled unamplified DNA the most.

Materials and methods: We analyzed the performance of PCR-based, multiple-displacement amplification (MDA)-based, and MDA-PCR combining WGA techniques (WGA kits Ampli1, REPLI-g, and PicoPlex, respectively) on single and pooled tumor cells obtained from EDTA- and CellSave-preserved blood and archival material. Amplified DNA underwent exome-Seq with the Illumina HiSeq2000 and ThermoFisher IonProton platforms.

Conclusion: We demonstrate the feasibility of single cell genotyping of differently preserved material, nevertheless, WGA and NGS approaches have to be chosen carefully depending on the study aims.

Keywords: CellSave; NGS; SNP; WGA; allelic dropout.

Figure 1. Distribution of identified known SNPs between datasets

(A) Known SNPs identified in single cells, amplified with Ampli1, PicoPlex, and REPLI-g WGA kits and obtained from EDTA-preserved blood in comparison to unamplified DNA. (B) Known SNPs identified in single cells, amplified with Ampli1 or PicoPlex and obtained from EDTA- and CellSave-preserved blood in comparison to unamplified DNA from unfixed cells. (C) Known SNPs identified in single CTCs, amplified with PicoPlex in comparison to each other.

Figure 2. Plots of CNA profiles along the whole genome (x axis)

(A) CNA profile of unamplified DNA from unfixed cells. (B–G) plots of CNAs in single SK-BR-3 cells, obtained from EDTA-preserved blood. (H, I) CNA profiles of individual CTCs, obtained from EDTA-preserved blood of the same breast cancer patient. WGA kits: (B, E) Ampli1; (C, F, H, I) PicoPlex; (D, G) REPLI-g.

Figure 3. Characteristics of pooled 1, 3, 5, and 10 SK-BR-3 cells, obtained from CellSave-preserved blood, amplified with Ampli1 and PicoPlex WGA kits, and sequenced with HiSeq2000 NGS platform

(A) Total identified SNPs. (B) Known identified SNPs. (C) Concordance of identified SNPs with reference dataset. (D) Sensitivity of the SNP calling analysis. (E) Allelic dropout. (F) Correlation of CNA profiles with CNA profile of unamplified DNA.