Next-Generation Sequencing: Reading Millions of Pieces at Once

From one read at a time to millions in parallel

Next-generation sequencing (NGS), also called high-throughput sequencing, changed the economics of genomics completely. Instead of reading one fragment at a time, NGS chops the genome into millions of short pieces and reads all of them at the same time, in parallel, on a single chip. Each short stretch of letters it produces is called a read — typically 100 to 300 letters long.

Short reads create a puzzle: millions of fragments with no labels saying where they belong. The fix is to align each read against the reference genome, like fitting puzzle pieces onto a finished picture. Because reads land randomly, every position gets covered several times over. That redundancy is called depth of coverage, and it is how sequencing tells a real letter from a machine error.

Reference: ...A C G T A C G T A C G T A C G T...
Read 1:    ...A C G T A C
Read 2:      C G T A C G T
Read 3:          T A C G T A C G
Read 4:              C G T A C G T A
             ^ each base read multiple times
30x coverage = each letter seen ~30 times
 -> a single odd read is outvoted as an error

Coverage means each base is read many times; agreement across reads separates true letters from noise.

Whole genome, or just the exome?

Whole-genome sequencing reads all 3.2 billion letters. But often you care most about the ~1–2% that codes for proteins. Sequencing just that part — the exome — captures most known disease-causing variants for a fraction of the cost and data. The trade-off: you miss everything in the non-coding regions, where many regulatory and disease signals also live.

Cheap NGS made population-scale projects possible. The 1000 Genomes Project sequenced thousands of people from around the world to catalogue normal human variation. That catalogue is what lets us tell, for any letter in your genome, whether your version is common or rare — the question at the heart of guides four and five.