Blended Length Genome Sequencing (blend-seq): Combining Short Reads with Low-Coverage Long Reads to Maximize Variant...

Area of Science:

• Genomics
• Bioinformatics

Background:

• Traditional short-read sequencing offers high accuracy but struggles with complex genomic regions and structural variants.
• High-coverage long-read sequencing improves variant detection but remains costly for routine single-sample analysis.

Purpose of the Study:

• Introduce blend-seq, a novel workflow integrating low-coverage long reads with standard short reads.
• Enhance single-sample variant discovery, including single nucleotide polymorphisms (SNPs) and structural variants (SVs).
• Improve variant phasing and genotyping accuracy cost-effectively.

Main Methods:

• Developed the blend-seq workflow combining 30x short-read data with 4x long-read data.
• Evaluated SNP discovery accuracy against varying short-read coverages.
• Assessed structural variant recall and precision using low-coverage long reads alone and in combination with short reads.
• Compared variant phasing performance between short-read-only and blend-seq approaches.

Main Results:

• Blend-seq with 4x long reads and 30x short reads surpassed 60x short-read-only performance for SNP discovery.
• Low-coverage long reads alone yielded a threefold increase in structural variant recall with maintained precision.
• Integrating short-read data improved genotyping accuracy for structural variants.
• Blend-seq significantly outperformed short-read phasing by leveraging long-context genomic information.

Conclusions:

• Blend-seq offers a cost-effective strategy to significantly improve variant discovery and phasing for single samples.
• The complementary strengths of short-read (depth) and long-read (context, resolution) sequencing are effectively leveraged.
• This approach advances genomic analysis by enhancing detection of SNPs, structural variants, and phasing accuracy.