Analysis of encrypted genomes

Mathematicians have turned the genome into a homomorphic cryptosystem

Microsoft cryptologists have developed an algorithm for encrypting DNA sequences, which allows them to be analyzed by traditional bioinformatic methods and at the same time does not compromise the owner of the genome. The technology was presented at the conference of the American Scientific Society AAAS, Science (How to Hide Your Genome) writes briefly about it.

The technology turns a DNA sequence into a homomorphic cryptosystem. The essence of such a system is that the user can manipulate data without decrypting it. In this case, the result of manipulation after decryption will be the same as if manipulations were carried out with uncoded information.

For the first time such a system was introduced in 2009 by IBM. It is based on mathematical lattices – a lattice in an n-dimensional vector space is a set of linear combinations with integer coefficients of vectors from a fixed basis. For the needs of cryptography, the problems of finding a vector of minimum length according to a given basis and the problems of finding the lattice vector closest to some arbitrary vector in space are used. These tasks are difficult to solve, however, it is quite simple to check the existing solution. Such one-sided complexity ensures the cryptographic stability of algorithms.

A two-dimensional lattice and vectors in it with two bases
Image: Oded Regev/Daniele Micciancio

Working with homomorphically encoded information requires much more computing power than manipulating open sequences. Therefore, in order to make the system practically applicable, scientists had to change a purely homomorphic algorithm and sacrifice the flexibility of calculations in favor of their speed. As an example of applicability, the authors claim that they were able to calculate the risk of a heart attack using a conventional personal computer based on the patient's genetic information in just 0.2 seconds. However, the nature of the genetic information used in the Science message is not specified.

Now anonymous but open DNA sequences are used in mass genetic research. In early 2013, Yaniv Erlich from the Whitehead Biomedical Institute in Cambridge showed that based on open genetic data, it is possible to identify the people to whom it belongs. It should be noted that this became possible due to the fact that data on the haplotype, as well as on the origin, age and place of life of those who were deanonymized by Ehrlich, were officially publicly available.

Portal "Eternal youth" http://vechnayamolodost.ru17.02.2014