Beijing Genomics Institute - Bioinformatics Technology

Bioinformatics Technology

De novo sequencing requires aligning billions of short strings of DNA sequence into a full genome, itself three billion base pairs long for humans.

BGI’s computational biologists developed the first successful algorithm, based on graph theory, for aligning billions of 25 to 75-base pair strings produced by next-generation sequencers, specifically Illumina’s Genome Analyzer, during de novo sequencing. The algorithm, called SOAPdenovo, can assemble a genome in two days and has been used to sequence an array of plant and animal genomes.

BGI’s 500-node supercomputer processes 10 terabytes of raw sequencing data every 24 hours from its current 30 or so Genome Analyzers from Illumina. The annual budget for the computer center is US$9 million.

SOAPdenovo is part of "Short Oligonucleotide Analysis Package" (SOAP), a suite of tools developed by BGI for de novo assembly of human-sized genomes, alignment, SNP detection, resequencing, indel finding, and structural variation analysis. Built for the Illumina sequencers' short reads, SOAPdenovo has been used to assemble multiple human genomes (identifying an eight kilobase insertion not detected by mapping to the human reference genome) and animals, like the giant panda.