Background

Bulk DNA sequencing, although being widely used nowadays, has been proven to be inadequate in analysis of heterogeneous systems, such as cancer tissues, which contain cancer cells of genetic aberrations at various degrees among normal cells with little or no genetic aberrations. Noncancerous cells can contribute a significant portion of the total DNA extracted from tumors, potentially masking important genetic aberrations (Alioto et al., 2015). Even when normal cells are removed, bulk sequencing of cancerous cells still averages out both the heterogeneity of cancerous cells in a tumor tissue and genomic instability over time (Yang et al., 2013; Francis et al., 2014). Single-cell DNA sequencing is believed to be the only method to reveal unequivocally the dynamics of mutations of tumor cell subpopulations in detail over space and time (Navin, 2015). Copy number variations (CNV) is under-detected in bulk sequencing, while they are found to be early events in tumorigenesis (Navin, 2015).

In the past years, several single-cell library preparation methods have been reported, which include DOP-PCR (Baslan and Hicks, 2014), MDA (Fan et al., 2011), MALBAC (Zong et al., 2012), LIANTI (Chen et al., 2017) and TnBC (Xi et al., 2017). Due to the fact that the minute amount of DNA from a single cell is not sufficient for NGS directly for most purposes, the single-cell genome needs to be amplified. Reflecting this requirement, almost all single-cell library preparation methodologies are named after an amplification method. As amplification is involved, the biases and errors associated with amplification inevitably need to be addressed (Xi, 2018). Biased amplification will require deeper sequencing to gain coverages. In an extreme case, under-amplified regions can be falsely identified as deletions. Amplification errors introduced by polymerases may overwhelm authentic mutations, which adds difficulty in mutation-calls. As TnBC methodology employs unique fragment index (UFI), it can handle amplification biases and polymerase-introduced errors better than other methods do (Xi et al., 2017; Xi, 2018).

An engineered Mu transposase was used in our original paper of TnBC library preparation (Xi et al., 2017). Since preparing custom-made transposases is technically demanding and time-consuming, here we report a protocol that utilizes Nextera, a commercially available Tn5 transposase. This protocol is intended to obtain single cell libraries that will be good for CNV detection through shallow sequencing. Due to the proof-reading DNA polymerase that is used in our library amplification, the error rate can be significantly smaller than that from DOP-PCR, MALBAC, or LIANTI. Therefore, the aggregate of sequences from multiple single cells can be used to detect global SNV of the source tissue of the single cells (Knouse et al., 2016).