ALBUQUERQUE, NM (GenomeWeb News) — Long-read sequencing can enable researchers to not only detect what drug-resistance mutations are present in someone with HIV, but also whether those mutations are present in one strain or spread across multiple strains of the virus circulating in that patient, according to a presentation at the Association of Biomolecular Resource Facilities annual conference, held here this week.
Dawei Huang, a supervisor at the DAVID Bioinformatics Lab of the US National Institute of Allergy and Infectious Diseases, noted in a talk at the ABRF conference that Sanger sequencing is currently used to detect drug resistance mutations in HIV, but the sensitivity of that approach limits its ability to perceive minor mutations. While he noted that next-generation sequencing improves upon that sensitivity to detect rare mutations, newer long-read approaches can detect both mutations and quasi-species of the HIV virus.
Applying these approaches both to a benchmark study and a patient case study, Huang and his colleagues were able to detect rare mutations in HIV. Additionally, by examining patient samples taken at different time points, they could determine how previously rare mutations became more common.
"We can clearly see the dynamic," Huang said.
While the number of people being infected with HIV each year has leveled off and death from AIDS has declined, there are some 30 million people living with HIV globally, and there is no cure.
Drug resistance and identifying those mutations that confer it are one challenge for successful treatment. Early detection of resistance — before it is apparent by viral loads — is needed, Huang said.
Most resistance mutations crop up in the stretch of HIV genome that houses its protease and reverse transcriptase genes. To detect drug resistance mutations, that 1.4 kilobase region is amplified using PCR and genotyped using the Sanger-based TruGene kit from Siemens, a method that has US Food and Drug Administration clearance. From the resulting chromogram, Huang said, mutations can be called.
But rare mutations — minor mutations that make up less than 20 percent of the calls — go undetected, he said.
Other approaches, such as next-generation sequencing approaches like Roche/454, Illumina's MiSeq, and Life Technologies' Ion PGM have a higher sensitivity for detecting these minor mutations. However, short-read approaches lose the linkage relationship between the mutations — it could detect multiple mutations, but not whether they were all in one strain or housed among a few strains circulating in the patient.
Siemens announced in late 2011 that it was partnering with Illumina to convert Siemens' Trugene HIV genotyping assay to run on the MiSeq.
A long-read approach, such as the Pacific Biosciences platform, Huang said, could give both the frequency and phasing information because its read length covers the full length of the PCR product. According to the company, reads longer than 10 kilobases are common, and it plans to increase average read length further this year.
In a benchmark study to evaluate the sensitivity of the three approaches for detecting rare mutations, Huang created admixtures of a wild type strain and a strain with 16 rare variants with each strain present at varying frequencies.
After PCR amplification of the viral region, the Sanger-based approach couldn't detect rare mutations below about 20 percent frequency; the next-gen sequencing approaches were sensitive down to about 1 percent; and the PacBio approach could detect them down to nearly 0.1 percent.
Huang and his colleagues also applied these three sequencing technologies to detect drug resistance mutations in blood samples taken from a patient at two time points. Those blood samples, Huang noted, were taken in January and February 2013, and in that time, the amount of virus in the patient's blood jumped dramatically from 3,000 copies per milliliter to 30,000 copies per milliliter.
In samples from the patient, the Sanger-based method could detect three dominant mutations in the January sample and six in the one from February. The NGS approach similarly could detect the mutations, but not how they arose.
The long-read approach, though, revealed that the three-mutation strain was the dominant one in January, but that it waned by February as the six-mutation strain — which it found to be present at about 0.4 percent in January — waxed to 91.85 percent by February to become the new dominant strain in the patient. The rare strain quickly developed as the wild-type strain declined.
Knowing which mutations are present and their phasing information can help clinicians decide upon a drug treatment regimen for the patient. Different drugs, Huang said, might be needed to target a virus strain with two mutations as compared to two strains with one mutation each.