The Sequencing Bead Array (SBA), a next-generation digital suspension array

Abstract

Here we describe the novel Sequencing Bead Array (SBA), a complete assay for molecular diagnostics and typing applications. SBA is a digital suspension array using Next-Generation Sequencing (NGS), to replace conventional optical readout platforms. The technology allows for reducing the number of instruments required in a laboratory setting, where the same NGS instrument could be employed from whole-genome and targeted sequencing to SBA broad-range biomarker detection and genotyping. As proof-of-concept, a model assay was designed that could distinguish ten Human Papillomavirus (HPV) genotypes associated with cervical cancer progression. SBA was used to genotype 20 cervical tumor samples and, when compared with amplicon pyrosequencing, was able to detect two additional co-infections due to increased sensitivity. We also introduce in-house software Sphix, enabling easy accessibility and interpretation of results. The technology offers a multi-parallel, rapid, robust, and scalable system that is readily adaptable for a multitude of microarray diagnostic and typing applications, e.g. genetic signatures, single nucleotide polymorphisms (SNPs), structural variations, and immunoassays. SBA has the potential to dramatically change the way we perform probe-based applications, and allow for a smooth transition towards the technology offered by genomic sequencing.

Figure 1. Schematic illustration of the Sequencing Bead Array (SBA) technology.

The digital reporter selection workflow is comprised of target hybridization, reporter activation, immobilization, selection, priming and reporter calling. (a) A synthetic oligonucleotide viewed from the bead 5´-end consists of i) a general linker site for bead coupling (black), ii) a target hybridization site (green), iii) a target unique ID trace (orange), and iv) a general priming site (purple). A reporter consists of a multifold of identical clustered oligonucleotides on a bead. Each reporter is assigned a unique ID trace corresponding to specific biomarker target of interest (e.g. X). A library of reporters includes multiple reporter types, that each targets a specific biomarker (e.g. X, Y, Z). (b) Biotinylated target molecules X (green) and Z (dark red) (c) Hybridization of target molecules activate corresponding reporters by the addition of a biotin molecule. Activated reporters are immobilized and are magnetically selected via biotin coupling to streptavidin-coated paramagnetic beads (brown). (d) The number of reporters passing through the magnetic selection renders digital information of biomarker presence. (e) A sequencing primer complementary to the priming site is added at high concentrations to saturate all beads and, due to the reporter-oligonucleotide clusters, each selected reporter amplifies its own signal. Sequencing of the ID trace identifies all reporters in parallel. (f) Data is processed and reporters are called and sorted by Sphix software. Due to unspecific interactions between the reporters and the magnetic beads, some background noise (un-selected reporters) will survive the assay. Therefore a threshold line (dashed black line) is included to differentiate a positive from negative call. In this case X and Z were detectable above the threshold line.

Figure 2. Reporter library construction and Sequencing Bead Array (SBA) assay performance.

In (a-c), the left-hand side bar-histograms detail y-axis average read counts, while the right-hand side detail y-axis read distribution percentages amongst the ten reporters. Error-bars are also included in the right-hand side graphs to illustrate the read distribution variability amongst the replicates. The upper error-bar value is the highest observed frequency, while the lower error-bar value is the lowest observed frequency. (a) Results from six individually constructed reporter libraries sequenced side-by-side without any prior selections (metadata Table S2). (b) Results from six parallel negative (water) selections showing the background noise effects (metadata Table S3). (c) Results from six quintuple HPV selections of equimolar amounts of PCR-products derived from Human Papillomavirus (HPV) plasmids HPV-16, HPV-18, HPV-33, HPV-45, and HPV-59 (metadata Table S4). (d) Read count bar-histograms (b and c) superimposed with the experimentally established threshold value (dashed line) of ~1261.5 reads, which clearly distinguishes a positive call from the background noise.

Figure 3. Sequencing Bead Array (SBA) assay screens of 20 cervical tumor samples.

Bubble chart representation of the SBA assay screens of 20 cervical tumor samples (extracted genomic DNA). The data is based on number of called reporters (metadata in Figure S1 and Table S5). Results were called as follows: ten single infections for Human Papillomavirus (HPV) genotype HPV-16, four single infections for HPV-18, two single infections for HPV-59, one single infection for HPV-45, one dual co-infection for HPV-16, and HPV-18, one dual co-infection for HPV-16, and HPV-45, and one genotype negative sample. OM-1299 was negative as no reporter read count was above the background noise threshold.