Table 1 A comparison among different sequencing technologies for structural variant detection
From: Towards accurate and reliable resolution of structural variants for clinical diagnosis
Platform | Read length | Cost | Comments | Run time |
|---|---|---|---|---|
Short reads (Illumina) | NovaSeq: up to 250 bp | $ | Short-read NGS performs well for >1kb regions. It struggles with shorter CNV detection 50-500bp, and in complex genome regions | NovaSeq: 0.15Tb/day |
10X Genomics Chromium | Up to ~100 kb | $$ | Sparse sequencing rather than true long reads; more complicated to align, with poorer resolution of locally repetitive sequences. However, 10X Genomics Chromium is currently discontinued | - |
PacBio SMRT sequencing | 10–15 kb (average) and up to 100 kb | $$$ | HiFi: long reads (10-20kbp) of high fidelity having a similar error rate as Illumina. CLR: Longer raw reads have high error rates dominated by false insertions; requires new alignment and error correction algorithms | 20 Gb/day |
Oxford Nanopore | averaging ~10 kb and up to 2 Mb | $$$ | Raw reads have ~5% error rates dominated by false deletions and homopolymer errors; often requires new alignment and error correction algorithms | A MinION Flow Cell : ~ 25 Gb/day |
Hi-C-based analysis | <100 bp | $$ | Sparse sequencing with highly variable genomic distance between pairs (1 kb to 1 Mb or longer); Detection may result from random chromosomal collisions Less than 1% of DNA fragments actually yield ligation products. Due to multiple steps, the method requires large amounts of starting material | Whole analysis within 28 hours |
BioNano Genomics optical mapping | ~250kb or longer | $ | Limited algorithms to discover high-confidence alignment between an optical map and a sequence assembly | 100x coverage of 3 human genomes is collected in less than 6 hours |