HiSeq X and NovaSeq Performance: Optimization for Clinical Applications

High-throughput parallel nucleotide sequencing has revolutionized genomic research and reshaped applications in clinical health care. The HiSeq X Ten and NovaSeq platforms further expand these opportunities with unprecedented... [ view full abstract ]

High-throughput parallel nucleotide sequencing has revolutionized genomic research and reshaped applications in clinical health care. The HiSeq X Ten and NovaSeq platforms further expand these opportunities with unprecedented capacity. The Human Genome Sequencing Center at Baylor College of Medicine recently acquired a NovaSeq 6000 and have 3+ years’ experience with the HiSeq X Ten system.  Here I will compare and contrast the two platforms with an eye to eventual deployment in the clinical environment.

To date, we have analyzed 3,917 HiSeq X flowcells, representing >31,234 30X human genomes. These studies have included primarily common disease cohorts, inherited cancers and mendelian disease. PCR-Free library methods were evaluated and implemented for optimized coverage in GC-rich regions. Metrics related to coverage, sample integrity and variant representation were established to ensure high quality genome sequencing. We have implemented standard metrics including % pass filter, % aligned bases, % error rate, % unique reads and % Q30 bases to achieve > 90 GB unique aligned bases per lane. Genome coverage metrics are also tracked for 90% of genome covered at 20x and 95% at 10x with a minimum of 86x10⁹ mapped, aligned bases with Q20 or higher. Additional metrics such as library insert size, duplicate reads, error rates, % pair reads and mean quality scores are also monitored. Platform sensitivity and precision at 30x coverage was determined to be 97.8% and 99.6% respectively using control sample NA12878.

When considering clinical WGS, costs for required coverage are paramount. High SNP, Indel and SV concordance between 20X and 30X WGS suggests that for variant calling 20X would be sufficient. However, the typical clinical exome benchmark is high with consistent exonic coverage of clinically relevant genes. Our analysis shows that clinical WGS may require ~40X coverage to reach this benchmark using HiSeq X.

While HGSC-BCM is able to produce high quality data with the HiSeq X platform, the platform is limited to WGS only.  The clinical field is moving swiftly towards WGS, but Exomes are still a widely-used option due to cost and time considerations.  This makes NovaSeq’s application flexibility and fast sequence generation an advantage. However, the NovaSeq platform should meet, if not, surpass the current high quality of the HiSeq X data in order to be considered for the clinical space.  HGSC-BCM acquired a NovaSeq 6000 in March 2017 and in the coming months will assess its performance in sequence quality and uniformity in coverage.