Scientists have increased the speed of DNA decoding hundreds of times.
Conventional DNA decryption algorithms rely on the idea that a large number of copies of a DNA strand can be broken down into many small, easily readable fragments that partially overlap with each other. But in this case, decryption requires a lot of computing resources.
The authors of the new work decided to remake the algorithm so that it does not require such high computing power. They created a mathematical theory by which a genome can be encoded as a set of frequently occurring sequences of several letter-nucleotides, rather than single units.
The new approach can make the decryption process faster and glue overlapping DNA fragments together. As a result, the authors used less computer memory for similar calculations.
Our approach works even if the source material contains up to 4% errors. Coupled with cheaper sequencing machines, this paves the way for the democratization of genetic analysis.
Bonnie Berger, professor at the Massachusetts Institute of Technology and one of the study authors.
During the experiment, the authors tried to decipher human DNA. The process of assembling the human genome took only 10 minutes and required about 10 gigabytes of RAM.