1. VarSeq Suite Device Information¶
1.1. Software Version¶
VarSeq Suite version 2.2.4 released 2021-10-26
UDI: (01) 00850033723007(11)211026(10)2.2.4
The VarSeq Suite is manufactured by:
EU Authorized Representative:
EUDAMED Single Registration Number (SRN): US-MF-000014206
1.3. Intended Use¶
The VarSeq Suite is a set of diagnostic analysis software used as a component of in vitro laboratory developed testing. The software automates tertiary analysis of variants found in gene panels, exomes and whole genomes. The software allows the identification and classification of causal variants, generated by NGS assays, to aid in making informed decisions, targeted diagnoses and targeted therapy applications in hereditary disorders and cancer. It is intended to be used by health care professionals and researchers.
1.4. Indications for Use¶
The VarSeq Suite is designed for use in research and laboratory settings for the analysis of human samples for the diagnosis of rare Mendelian diseases or analysis of the molecular profile of cancers for prognostic, diagnostic and therapeutic guidance. The software is designed to be used by trained medical professionals. The analysis and resulting interpretations and reports are designed to be interpreted by physicians, oncologists and genetic counselors or other trained medical professionals, not patients. The software is designed to run inside the infrastructure of the clinical laboratories, without the need to transmit patient or genomic data outside of this environment. The software starts with the input NGS aligned reads and variant calls, and is not a device for collecting or preparing samples for sequencing or performing primary sequence analysis of base-calling from raw instrument data.
There are no specific contraindications associated with the VarSeq Suite.
The VarSeq Suite is not designed for use for personal genomics, the diagnosis or variant interpretation of healthy individuals or to be used by non-medically trained individuals.
VarSeq Suite is designed with flexibility as a core component, requiring configuration and customization to a given test environment. Following these precautions will reduce potential risks associated with implementing software that is part of a laboratory testing process:
- Customize and calibrate the provided starting workflows (project templates) based on representative samples processed entirely within the full set of test procedures, with QC thresholds and analysis parameters determined through the validation procedure for a given genetic test.
Validate the complete sample preparation, sequencing, secondary analysis and tertiary analysis pipeline
Define test-specific annotations and quality filters based on the levels of detection and analytical outcomes determined by the analysis of benchmark samples with variants present at known input quantitative characteristics.
Ensure that a backup and recovery policy is implemented that includes backing up of project data, user data and any input data for the software, to reduce risk of total data loss through hardware or system software level failures.
Plan hardware capacity of RAM and disk resources allocated to systems by performing benchmarking of production workflows with representative production samples to avoid undefined or terminating behavior caused system resource exhaustion. See section 1.9 System Requirements.
Implement anti-malware, system firewalls and other security best practices (see section 1.10 Cybersecurity Guidance) to avoid cybersecurity breaches at the system level that may allow read-level access to the software inputs, outputs and user data
Do not run the software or server components on directly accessible internet-facing hosts and ensure physical and system-level access is limited to authorized individuals to prevent unauthorized access through external networks or physical intrusions.
The VarSeq Suite is a decision support software in vitro medical device designed to aid in the diagnosis of rare diseases as well as aid in the reporting of the potential prognostic and therapeutic implications of genomic biomarkers inferred from the detection of genomic mutations (variants) using Next Generation Sequencing (NGS) instruments. Laboratories choose one of many NGS-based assay products to sequence all or a subset of the human genome in a select set of genes (gene panels) or all genes (exomes or genomes) that results in a set of raw variant calls for a given patient. The VarSeq Suite supports:
Reading input in industry standard VCF files containing small variants (detectable using short-read NGS data) and CNVs (detected by coverage profile analysis of NGS data)
Annotating imported data with numerous public and licensed databases
Evaluating sample and variant-level quality with metrics, derived statistics and supporting visualizations
Running algorithms to make in silico analyses of the variants impact on the patient’s genes and their functions
Following of industry standard guidelines for scoring and classifying germline and somatic small variants and CNVS through guided workflows and auto-scoring algorithms
Reviewing relevant clinical evidence and literature to assess whether a variant meets criteria for reported
Composing all selected variants as well as patient and sample level data into a customizable clinical report
Storing previously sequenced sample’s variants into a warehouse to compute cohort level statistics and per-variant frequencies to inform future analyses.
The product has several modules that serve the diversity of needs of the market, but is developed, deployed, and distributed as a single software solution: * VarSeq (Core): Import raw data, enhance with annotations and algorithms, visualize and review for quality, filter and prioritize for interpretation * VS-CNV: A module for handling a special type of variant: Copy Number Variants that require specialized detection, quality assurance and interpretation strategies * VSClinical: Implementation of two separate industry standard guidelines (for inherited vs somatic variants) that implement decision support for interpreting variants in the context of clinical testing * VSPipeline: A command line tool to automate the construction of projects containing any of the above capabilities in a high-throughput and automated environment. * VSWarehouse: A server installation that includes specialized database for storing and retrieving internal knowledge bases of previous interpretations and a store of variant data from all samples over time.
1.9. System Requirements¶
The VarSeq suite desktop app runs on the following 64-bit operating systems:
Linux Ubuntu 18.04 or 20.04 or RHEL/CentOS 7 or 8
Mac OS 10.13 or later (tested up to mac OS 11)
The following hardware requirements are minimally required per concurrent user of a system:
4 GB of RAM
2 CPU Cores
100GB of space available for annotations and projects
If analyzing genomes, exomes and large gene panels, the recommend requirements are:
16GB of RAM
8+ CPU Cores
1TB of space available for annotations and projects
1.10. Cybersecurity Guidance¶
As on-premise software, the establishment of a secure network and system environment is critical for protecting the sample, genomic and imported patient information analyzed by the VarSeq Suite. In general, there should be limited or no direct access from the internet to the hosts running the VarSeq Suite and the storage system containing input and project data.
Note: While VarSeq Suite does not require any patient personal identifiable information to be imported into the system, it is common to include as part of a laboratory workflow. Additional per-sample information may be imported and placed on the corresponding generated report.
The following best practices are recommended to minimize cybersecurity risk:
Install institutional firewalls preventing unauthorized incoming connections
Install anti-malware and endpoint security software to prevent system level exploits that enable local system access and thus access to the locally installed software and data.
Back up all input, project and user data to non-network attached mediums to enable recovery from ransomware or total loss of data
Provide system level user authentication with a strong password policy and potentially two-factor authentication to ensure access is restricted to the installed software and data
Ensure physical level security preventing un-authorized access to workstations. Include auto-locking on inactivity policies to ensure unattended workstations cannot be accessed by unauthorized personnel
Contact your network or IT administrator or specialist to review and implement these best practices in accordance with your institutional policies.
To be compliant with GDPR Privacy rights while using the VarSeq suite, we suggest the following implementation policies:
Create one project per sample. The allows for precision when complying with a request to delete the personal data of a test subject.
While a project is retained, updates or rework of personal data can be done from within VSClinical and a new report can be generated
While a project is retained, a copy of all personal data can be exported using standard file formats
Do not upload personal data to VSWarehouse, instead choose to drop all sample-level fields during the upload process and send only sample tracking identifiers.
Note: While guidance above addresses the ability for VarSeq Suite to be used in a GDPR compliant manner, it is ultimately the responsibility of the lab processing and storing personal information to ensure compliance GDPR. Please see the GDPR Requirements for the full list of responsibilities, policies, and procedures a lab must implement.
1.11. Maintenance and Upgrade Guidance¶
The VarSeq Suite is designed to be validated with a specific version of the software and update in coordination with the cycle of re-validation of laboratory processes. No updates are required in-between these updates, although notification of releases notes and known issues should be monitored to ensure that there are no issues that impact production workflows. Similarly, annotations used during analysis may be locked to the exact versions used during validation, although it is common to stay up to date with monthly updated annotation sources such as ClinVar and OMIM, the impact of these regular changes must be evaluated in terms of the test provided.
When choosing to update the software, all production workflows should be re-validated with the same benchmark samples used in the initial calibration and validation step and changes to results matched to expected changes such as updated annotations, algorithm improvements and new entries in lab-specific knowledge-base that impact analysis.
Standard system monitoring and maintenance should be performed for the hosts running the VarSeq Suite, including monitoring for the volumes containing any user data, shared application data or project data running out of capacity or experiencing degraded performance. Security measures and system software patches should be maintained and kept up to date to reduce vulnerability exploitation risk.
1.12. Analytical Performance¶
The VarSeq Suite has the analytical performance evaluated with the following automated assessments:
Comparison of the VSClinical ACMG Classifier classification in directional agreement with ClinVar
Comparison of the VSClinical ACMG Classifier recommended criteria in agreement with ClinGen expert curations teams manually scored criteria for automatable criteria and genes without specialized scoring systems
Comparison of VS-CNV on 100 gene panels that had known MLPA results
The results of these comparisons demonstrate analytical validity.
1.13. Clinical Performance & Risk/Benefit Information¶
The VarSeq Suite has the clinical performance evaluated with studies conducted with users performing clinical variant interpretation in their own laboratory setting. These studies demonstrated that the VarSeq Suite improves the efficiency (time to completion) of performing variant scoring by automating analysis and providing workflows for decision support over the alternative of manual variant interpretation and reporting. This improved performance with no measured drop in quality as measured by unexplained discordance in outcomes outweighs the risk of using automated software.