Recommendations for bioinformatics in clinical practice

Abstract

Background: Next-generation sequencing (NGS) is well established in clinical diagnostics, and whole-genome sequencing (WGS) is increasingly becoming the method of choice, as a result of lower prices and robust comprehensive data. While guidelines exist for variant interpretation and laboratory quality considerations, there remains a need for standardised bioinformatics practices to ensure clinical consensus, accuracy, reproducibility and comparability.

Methods: This article presents consensus recommendations developed by 13 clinical bioinformatics units participating in the Nordic Alliance for Clinical Genomics (NACG) by expert bioinformaticians working in clinical production. The recommendations are based on clinical practice and focus on analysis types, test and validation, standardisation and accreditation, as well as core competencies and technical management required for clinical bioinformatics operations.

Results: Key recommendations include adopting the hg38 genome build as reference, and a standard set of recommended analyses, including the use of multiple tools for structural variant (SV) calling and in-house data sets for filtering recurrent calls. Clinical bioinformatics in production should operate at standards similar to ISO 15189, utilising off-grid clinical-grade high-performance computing systems, standardised file formats and strict version control. Reproducibility should be ensured through containerised software environments. Pipelines must be documented and tested for accuracy and reproducibility, minimally covering unit, integration and end-to-end testing. Standard truth sets such as GIAB and SEQC2 for germline and somatic variant calling, respectively, should be supplemented by recall testing of real human samples that have been previously tested using a validated method. Data integrity must be verified using file hashing, while sample identity must be confirmed through fingerprinting and genetically inferred identification markers such as sex and relatedness. Finally, clinical bioinformatics should encompass diverse skills, including software development, data management, quality assurance and domain expertise in human genetics.

Conclusions: These recommendations provide a consensus framework for standardising bioinformatics practices across clinical WGS applications and can serve as a practical guide to facilities that are new to large-scale sequencing-based diagnostics, or as a reference for those who already run high-volume clinical production using NGS.

Fig. 1

Overview of the NACG community. A Geographic distribution of NACG sites: a map illustrating the 12 member sites of NACG network and one guest site from the Netherlands. B Annual analysis volume per site: the volume of analyses performed per year across NACG sites. The x-axis shows four categories of analysis volume, while the y-axis indicates the number of sites (out of 13) in each category. Different colours represent various types of analyses, as indicated by the legend. C Genome reference build usage: a summary of genome reference build statistics across the 13 participant sites. D Details of pipelines in clinical production: the y-axis shows the type of analysis, and the x-axis shows the type of variant or metric reported. Bubble size reflects the number of sites in each intersection category. The colour legend is the same as in panel B

Fig. 2

Summary of reported coverage and quality threshold used in clinical production. Consensus of minimal thresholds and quality metrics remains limited for targeted panels and WES, whereas a minimum sequence coverage depth is agreed upon across all units for both somatic and germline testing

Fig. 3

Reported use of test concepts to ensure quality of pipelines. The typical types of tests and test phases are shown on a polar barplot with the following notation for testing: Unit, testing individual components or functions in isolation; Integration, testing the interaction and integration between multiple components or modules; System, testing the entire system as a whole to ensure it meets the specified requirements; Regression, re-testing previously working functionality to ensure that changes have not introduced new issues; End-to-end, verifying the flow and functionality of the entire system from start to finish; Performance, evaluating the system’s performance and responsiveness under different workloads and scenarios; Security, identifying vulnerabilities and ensuring the system’s resistance against potential security threats; Usability, assessing the system’s user-friendliness and ease of use. The preferred test type overall is shown in grey. Counts on the barplot indicate the number of sites using each test type at each stage