A Software Tool for Reagent Design to Expand Access to Single-Nucleotide Variant Detection by the Oligonucleotide Ligation Assay

Abstract

Single-nucleotide variants (SNVs) and polymorphisms are characteristic biomarkers in various biological contexts, including pathogen drug resistances and human diseases. Tools that lower the implementation barrier of molecular SNV detection methods would provide greater leverage of the expanding single-nucleotide polymorphism/SNV database. The oligonucleotide ligation assay (OLA) is a highly specific means for detection of known SNVs and is especially powerful when coupled with PCR. Yet, the OLA design process remains intensive, and criteria for success are uncertain. To assist in the design process, this study describes OLAgen, an open-source tool to automate development of OLAs and their coupled PCR assays. The software facilitates alignment of sequences surrounding SNVs and generates ligation probes while screening for dimerization potential. OLAgen successfully produced ligation probes that closely matched previously validated designs for HIV-1, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and KRAS, confirming its reliability and potential for clinical applications. The tool was used to generate new assays targeting Mycobacterium tuberculosis drug resistance and variants in the human JAK2, BRAF, and factor V genes, all of which demonstrated 100% sensitivity and specificity in controlled laboratory experiments. The OLAgen predicted assay designs detected mutant frequencies as low as 1% to 5% in wild-type backgrounds in proof-of-concept laboratory studies. OLAgen represents a significant advancement in accessible assay design, promoting the broader application of OLA technology in clinical and research settings.

Figure 1

Contiguous product generated during the oligonucleotide ligation assay (OLA) is sensitively detected by PCR. Left: The ligation event occurs if and only if the 3′-base of the variable probe complements the target single-nucleotide variation. Right: The ligation joins the two synthesized strands into a contiguous DNA strand that contains two priming regions (blue) and a hydrolysis probe complementary region (red) for downstream detection by PCR. A, adenine; C, cytosine; F, fluorophore; G, guanine; Q, quencher.

Figure 2

The OLAgen tool designs oligonucleotide ligation assay (OLA) and subsequent PCR reagents. The design process undergoes five main steps. (1) The tool processes the user input file (FASTA) and recognizes the various sequences within the input. (2) OLAgen then aligns the recognized sequences to determine single-nucleotide polymorphisms (SNPs) between a reference sequence and target sequences. (3) Variable probe (VP) and common probe (CP) sequences are then generated to detect the determined SNPs. This step also checks for dimerization that may cause false ligation. (4) If the user wishes to couple OLA to PCR, OLAgen will work through a pre-established primer and hydrolysis probe (Hydr. Probe) library for candidates for the combined assay. (5) Full reagents for OLA and PCR are generated and provided to the user for export. ΔG, Gibbs free energy; RBD, receptor-binding domain.

Figure 3

The OLAgen software uses a linear and intuitive pipeline for reagent design. The simple OLAgen workflow works stepwise from start to finish (green, rounded rectangles) to develop oligonucleotide ligation assay (OLA) reagents and, if desired, PCR reagents. The OLAgen system automates burdensome design tasks (gray rectangles) with user feedback and engagement (blue diamonds) that enables greater breadth for particular use cases. OLAgen is dependent on predesigned tools and libraries (gray, double-bar rectangles). Seq, sequence; SNP, single-nucleotide polymorphism.

Figure 4

The OLAgen graphical user interface is simple and streamlined for maximal user friendliness. The standard workflow of OLAgen progress through three primary windows: (1) a main window to begin user submission or read help documentation, (2) a window to select reference targets and alignment type, and (3) a tab-facilitated window to manage output sequences for user export.

Figure 5

OLAgen-generated ligation assays exhibited dynamic sensitivity to mutant frequencies in wild-type background. Proof-of-concept variant frequency detection was performed for the four prospective ligation assays using oligonucleotide ligation assay–PCR across five mutant (MUT) frequencies in wild-type background [1% (red), 5% (orange), 10% (yellow), 50% (green), and 100% (blue)]. All assays demonstrated sensitivity to low-frequency variants and produced robust responses to changes in MUT frequency. Each assay exhibited a strong, predictive log-linear relationship (black line; shading = 95% CI) between variant frequency and quantification cycle (Cq) values, indicating consistent assay performance across a range of target concentrations. All sample data are presented and shown with a jitter in the x value for visual clarity. No target controls are not shown. n = 6 replicates per percentage MUT; n = 30 samples per assay.

Supplemental Figure S1

All files in OLAgen are encapsulated using model-view-controller (MVC) architecture. MVC is a design pattern that separates OLAgen's application logic, user interface, and data management to promote codebase organization, enhanced program scalability, and simplified maintenance. The application entry point, OLAgen_main.py, initializes the controller, which handles the program's application logic. The controller responds to user actions from the view, OLAgen's dynamic user interface, and commands the view to update when necessary. Simultaneously, the controller exchanges data with the model, ensuring that system inputs and outputs are properly stored and processed. The controller also calls upon utility files for specialized tasks, like screening primers and aligning sequences, generating a streamlined workflow during ligation-assay component generation.

Supplemental Figure S2

OLAgen uses the model-view-controller (MVC) architecture throughout its main programming files to streamline genetic data analysis and generation of ligation-assay components. OLAgen_main.py initializes the application by activating main_window.py, the primary controller file. Each controller file has access to its own user interface (ui) file, interface logic, and the overall system model and utility files. User interactions with main_window.ui prompt its corresponding controller to call upon fasta_window.py. Here, users upload a.fasta file for sequence analysis, allowing for the generation of the finalized ligation-assay components in output_window.py. The model file global_state.py serves as the application's main way to store and access data, whereas utility files like alignmentSource.py and olagenprocess.py provide specialized data analysis functions. Together, OLAgen's MVC file structure allows for better workflow visualization, programming organization, and project scalability, laying a strong foundation for future enhancements and updates.

Supplemental Figure S3

The four newly validated OLAgen assay designs fulfilled the criteria for success. PCR curve data for all individual samples are shown. Columns represent the oligonucleotide target in the data plot, and rows represent the assay type tested. Data were collected for each assay across two or three independent experiments. Each experiment contained equal counts of mutant, wild-type, and no target present samples tested. All data were normalized and baseline shifted. n = 20 per target; n = 60 per assay.

Supplemental Figure S4

OLAgen-generated ligation assays exhibited dynamic sensitivity to mutant concentrations in wild-type background. Proof-of-concept variant frequency detection was performed for the four prospective ligation assays using oligonucleotide ligation assay–PCR across five mutant (MUT) frequencies in wild-type background that correspond to various mutant concentrations [1 × 10⁴ copies/μL (red), 5 × 10⁴ copies/μL (orange), 1 × 10⁵ copies/μL (yellow), 5 × 10⁵ copies/μL (green), and 1 × 10⁶ copies/μL (blue)]. All assays produced robust responses to changes in MUT concentration. Each assay exhibited a strong, predictive log-linear relationship (black line; shading = 95% CI) between mutant concentration (shown here in log₂) and quantification cycle (Cq) values, indicating consistent assay performance across a range of target concentrations. All sample data are presented. No target controls are not shown. n = 6 replicates per percentage MUT; n = 30 samples per assay.