One of the goals of genetic studies is to characterize genetic variation in individuals with specific conditions in order to identify variants associated with disease or efficacy of treatment modalities. Recently, massively parallel sequencing technology has made it possible for an individual's genome to be examined in fine detail. The increased use of this technology, often called next-generation (NGS) or deep sequencing, paired with powerful bioinformatic analyses of the resulting data, has facilitated the identification of novel disease-causing variants. Targeted sequencing of the genome's coding regions has been used to identify genes associated with rare monogenic disease including Kabuki syndrome , familial amyotrophic lateral sclerosis (ALS) , Miller syndrome  and Van Den-Ende-Gupta syndrome . Currently, large sequencing projects, such as the 1000 Genomes project (http://www.1000genomes.org/) , are using this technology to characterize human genome variation on a population-based scale. As the cost of deep sequencing continues to decrease, the use of NGS technology will surely increase.
As deep sequencing projects are completed, additional DNA from study participants will be needed for replication and follow-up studies. While DNA derived from a subject's peripheral whole blood is a preferred source of starting genetic material, continued access to the participant for additional venipuncture may not be possible, or DNA isolated from peripheral whole blood may be available in limited quantities. Given these limitations, lymphoblastoid cell lines (LCLs) provide a convenient alternative. LCLs, created through the in vitro infection of B-lymphocytes with the Epstein-Barr virus (EBV), can provide an unlimited and lasting resource of the patient's genetic material. LCLs are well suited for many types of studies including genome-wide association [6, 7], functional genomics , proteomics  and pharmacogenomics [10, 11]. Furthermore, LCLs and their DNA can be made available to many investigators worldwide through biorepositories [12, 13].
Despite the frequent use of LCL for biological research, concerns have been raised regarding potential genomic changes that may be introduced during cellular transformation and subsequent cell culturing. Several investigations have addressed this issue. For example, DNA copy number changes have been detected following extensive passaging of cell cultures . The fidelity of genotype calls between DNA derived from LCLs and PBMCs from the same individual also has been examined [15–17]. These studies used gene chips to compare genotypes between the paired samples. Even though no significant changes were observed, this approach only interrogated the SNPs represented on the chips. Newly induced mutations may be introduced during the creation of the LCLs and/or after subsequent expansion of the derived cell lines. Recent studies have highlighted the association of de novo mutations with common disorders such as autism , schizophrenia  and mental retardation . Therefore, determining if these mutations are real or an artifact of the starting material is of great importance as false-positive results can be introduced into the study design.
Recently, within the 1000 Genomes Project, the presence of de novo mutations in two trio families was described. The authors estimated that 0.61% of coding variants identified were de novo . Since this study used DNA derived from LCLs, they were unable to compare the results to DNA derived from PBMCs in order to determine if these de novo mutations are real or induced through the cell transformation and culturing process.
The aim of the present study was to determine if DNA from EBV-transformed B-lymphocytes contains new mutations when compared to DNA from untransformed material. To address this, we performed whole exome-sequencing using both PBMC- and LCL-derived genomic DNA from a family of 4 individuals.