How does virco®TYPE work?
virco®TYPE HIV-1 is a refined HIV-1 resistance analysis test that provides a combination of genotypic and phenotypic information, complemented with additional clinical information. The test results are interpreted in a clinical context with Clinical Cut-Off values to ensure maximal utility for making therapeutic decision in daily practice.
The biggest difference with other tests that provide phenotypic information is that virco®TYPE HIV-1 does not rely on a complicated, time consuming and labor-intensive assay to produce a Fold Change in IC50 result. Instead, it uses a sophisticated bio-informatics technology, Virtual Phenotype™- LM to calculate phenotypic resistance from the genotype.
The phenotype predictions are based on a large and continuously growing database of previously determined Genotypes and Phenotypes.
1. Databases
Virco has established and maintains several databases with critical information that is used to make calculate the phenotype from the genotype, and to develop Clinical Cut-Off values for interpreting the calculated phenotype.
A. The Genotype - Phenotype Correlative Database forms the basis of the phenotypic calculations and consists of previously determined genotypes and matched phenotypes from clinical isolates. This database is continuously updated to reflect the most up to date HIV epidemiology and grows on average by 20% per year for most drugs. This growth is more significant for newer drugs.
- Genotypes: Virco's genotyping database contains 445,000 (August 2010) genotypes from clinical isolates obtained from patients across the world, and is thus a valuable resource to study the prevalence and patterns of resistance. Please consult the Congress Presentations section for examples of such research projects.
- Phenotypes: Like the genotyping database, the phenotyping database contains data from 98,000 (August 2010) clinical isolates. All phenotypes were determined using Virco's Antivirogram® assay, and for each sample the database contains Fold Change in IC50 measurements for all drugs that were available at the time the assay was performed.
- Correlative Database: The correlative database is the intersection between the genotyping and the phenotyping databases and contains greater than 61,000 samples (August 2010) for which both a genotype AND a phenotype are available. The correlative genotype-phenotype database is the basis for the calculations in virco®TYPE HIV-1.
B. The Clinical Outcomes Database contains clinical data from 16,074 (August 2010) patient records. This includes treatment history, current treatment, change in baseline Viral Load and CD4 count after the resistance test etc. The Clinical Outcomes database forms the basis for the virco®TYPE HIV-1 Clinical Cut-Offs. The Clinical Outcomes database forms the basis for the virco®TYPE HIV-1 Clinical Cut-Offs.
2. VirtualPhenotype
Virtual Phenotype™ is the name of Virco's proprietary system to make accurate phenotype calculations from the genotype. The first generation Virtual Phenotype™ was introduced in 2000, and replaced with a new and more accurate system, Virtual Phenotype™ - LM, in the spring of 2005. LM stands for Linear Modeling, the statistical technology underlying the phenotype predictions.
A. Virtual Phenotype™ - first generation
The first generation Virtual Phenotype™ used patterns of resistance-associated mutations to search the genotype - phenotype correlative database for viruses with similar mutational profiles. The average Fold Change of all database matches for the patient's virus was then calculated and reported.
B. Virtual Phenotype™ - LM
Virtual Phenotype™ - LM is the bioinformatics engine employed to fully utilize the extensive resistance information contained in Virco's coorrelative database. As of July 2007, the correlative database contained over 53,000 clinical isolates with both a genotype and a conventional measured (Antivirogram) on the same sample.
Virtual PhenotypeTM LM uses multiple linear regression modeling to identify individual mutations and pairs of mutations in the genotype of a virus that should be considerd in making an accurate calculation of fold-change. The model also assigns résistance weight factors (RWFs) for all the identified mutations and pairs identified. These RWFs are a measure of how much a given mutation or pair of mutations affects the fold- change calculation. RWFs can be either positive, resulting in higher fold change values, or negative, lowering FC as a result of the resensitizing effects of some mutations.
Unlike the first generation of the Virtual Phenotype that sought to identify matches to viruses with very similar mutational profiles, the Virtual PhenotypeTM LM calculates the fold change for every virus by combining the RWFs for mutations and pairs of muations that are identified in the model. Accurate fold-change values can be reported for every virus regardless of whether or not there are similar viruses in the database.
The multiple linear regression model is updated every two to three months as additional new genotypes and phenotypes are added to the correlative database. As compared to genotypic algorithyms , Virtual Phenotype LM incorporates many more mutations and pairs of mutations in assessing drug resistance. Because the model is fully data driven and objective, subjective rules that may not be up to date are not required.
To learn more about how VirtualPhenotypeTM - LM works, Please click here.
Compared to mutation based algorithms for genotype interpretation, vircoTYPE HIV-1 has the following advantages:
- Quantitative result that help to optimize the choice of drugs with partial activity
- Inclusion of more mutations than other genotypic systems for a complete picture of resistance
- Frequent database updates for an up-to-date resistance analysis
- Fully data-driven data analysis, taking into account weighted contribution of mutations for a realistic assessment of resistance
- Consistent Clinical Cut-Offs for a clinically relevant interpretation of the test results