What to look for in genome analysis software for clinical labs

To build or buy? That is a question that many labs face when it comes to finding a bioinformatics solution to support the analysis of next-gen sequence (NGS) data for clinical genetic testing.

?

Many of the larger volume, specialized genetic testing labs have opted to build their own genome analysis platforms in house. This affords them the ability to customize the platform to their exact needs and maintain control over feature development. It also provides a sense of stability and security for what may be the most critical component of their test.

?

However, building one’s own bioinformatics pipeline from scratch comes at a both a financial cost and time cost and may not be a feasible or desirable option for smaller volume labs or those new to NGS. Fortunately for these labs there are several good commercial solutions that provide genomics software as a service.

?

While using a commercial software solution may save time and resources, the decision to buy or build your own may come down to how robust, user-friendly, flexible and cost-effective commercial platforms are. Every commercial platform has its pluses and minuses, and the features required by different labs may vary depending on what tests they are offering.?

Desirable features for different use cases

None of the commercially available platforms are perfect for all intended uses, but some software platforms are better than others for specific use cases. Here are a few of the features that we found to be helpful for a few of the most common clinical testing scenarios.

?

High volume routine gene panel testing

Labs that offer gene panel tests, like those for hereditary cancers or cardiac conditions, carrier testing or newborn sequencing, want a scalable solution. They are looking for software platforms that allow them to analyze smaller single nucleotide variants (SNVs), and sometimes larger variants, efficiently and accurately.

?

Seamless import of sequence data and automated quality control can help streamline production on the front end. On the back end, automated report generation and export into both PDF and structured data make for a smooth experience.

?

High volume labs would also benefit from a platform that has a robust case management system allowing for tracking of the status of each case as it moves through the queue.

?

Most platforms employ some degree of automated SNV interpretation to improve efficiency, but automation needs to consider that the interpretation guidelines have evolved since 2015. Current ClinGen recommendations include up- and down-grading of criteria as well as some gene-specific thresholds. The next guideline version will use a completely different point-based system. Each lab will have its own SOP for variant classification. Therefore, it’s critical that the software platform is flexible, allowing for modification of the strength of criteria used for interpretation, and ideally the ability to apply gene-specific modifications.

?

Efficiencies can also be accrued over time if there is a way to access previously classified variant. Most platforms offer this feature, but nice to have add-ons include flagging older variants that would benefit from reclassification, automatic submissions to ClinVar, and a way to keep variant classifications private and not share with other users of the platform.

?

Whole genome/exome sequencing for rare disease diagnosis

With increasing reimbursement, using whole exome sequencing (WES) or even whole genome sequencing (WGS) to identify causal genetic variants in patients with rare genetic disorders becomes an attractive proposition for an NGS lab. Achieving the best diagnostic yield is key for these tests. Additionally, in some applications like rapid sequencing in the acute neonatal intensive care unit (NICU) setting, speed is of the essence.

?

Labs that offer WGS/WES are looking for software that allows them to analyze a wide range of variant types from SNVs, copy number variants (CNVs), other structural variants (SVs), Mitochondrial DNA (MtDNA) and repeat expansions. Automated variant classification is less of a priority.

?

Instead, these labs want robust annotations (such as databases of CNVs/SVs and MtDNA variants), flexible filtering options and tools to assist in identifying genes that match a patient’s phenotype.

?

Gene prioritization algorithms can dramatically improve the speed and efficiency of finding causal variants and are recommended, especially for labs wanting to offer rapid genome analysis.

?

Case interpretation may be divided into distinct steps that are performed by different individuals in the lab. As such, having a platform that supports custom workflows, where cases can be distributed and transitioned between individuals, will benefit the user.

?

Other scenarios

Labs deploying NGS may want to offer other types of genomic tests, from oncology-focused to polygenic risk scores (PRS) to pharmacogenomics, each of which requires its own set of features.?

?

Somatic NGS-based tests in cancer require different database annotations, variant classification tools, and customizable report templates. To accommodate scores like tumor mutational burden (TMB) based on the number of variants in a gene and not a specific variant, additional tools and reporting features are needed.

?

Similarly, labs that want to offer polygenic risk scores (PRS) need a way to combine genotypes across hundreds to thousands of specific variants and calculate a risk score based on that information.

?

Pharmacogenomic testing requires a secondary analysis pipeline capable of discerning haplotypes to call star alleles/diplotypes. Furthermore, the software needs to be able to annotate patient diplotypes with drug-related information from a custom database.

Questions to ask when evaluating a software solution

Below we outline some of the considerations when choosing a software solution for analysis of NGS data, or for designing your own software in-house.

Commercial platforms offer pricing models based on either the number of cases, the number of variants or the number of users. There are other costs as well. Some platforms require purchasing a one-time software license. Some require a minimum volume of cases or time commitment.

?

Ultimately, labs need to weigh the cost-effectiveness of building versus buying, depending on the expected throughput and the specific types of tests intended to offer. Fortunately, there are over a dozen options available today, making NGS more accessible to labs of all sizes.

?