Nature’s script for genetic information could be a way to keep things secret, says S.Ananthanarayanan.

An important part of commercial transactions is to control who can read a message, and then to assure the reader that the message is genuine. For centuries, the device to do this was the sealed envelope and the sender’s signature. With transactions over the Internet, however, there is a need for a different way to seal the message. And then to know that it has not been tampered with.

Cryptography, or the science of coding, was first of interest in the military and the world of espionage, and many ingenious methods have been devised. Not all of these, however, are useful in electronic communication of the rising numbers of banking transactions. The methods used here are automated coding, and they depend, largely, on a secret number, of many digits, which the sender and the receiver use to code and decode a message.

The core of secure transactions thus reduces to generating sufficiently complex numbers, that is to say, numbers that are difficult to guess, and to create them speedily and easily. This, however, can be a challenging thing to do, and several schemes, including those that use the quantum effects in lasers, have been developed. In a paper in the journal, Nature Communications, Linda C. Meiser, Julian Koch, Philipp L. Antkowiak, Wendelin J. Stark, Reinhard Heckel and Robert N. Grass, from the Swiss Federal Institute of Technology, Zurich, and the Technical University of Munich, describe a method based on the ways the DNA molecule assembles itself, which turns out to be effective indeed.

The relevant data, or the payload, of an electronic transaction is usually the identity of a paying and a receiving account and then the amount of transfer. Else, it could be the text of a contract, or the offer, or the acceptance, on the way to closing a contract. We can see that all of commerce hinges on assuring the secrecy and the integrity of such messages.

Electronic transmission of data is in the form of binary numbers, which consist of a series of ‘0’s and ‘1’s. For instance, the sentence, “Thirty days hath September”, reads like this in binary: “01010100 01101000 01101001 01110010 01110100 01111001 00100000 01100100 01100001 01111001 01110011 00100000 01101000 01100001 01110100 01101000 00100000 01010011 01100101 01110000 01110100 01100101 01101101 01100010 01100101 01110010”, with each group of eight numbers standing for an alphabet. Now, if this collection of ‘0’s and ‘1’s were jumbled up, the groups of eight numbers would stand for different letters, even numbers and punctuation marks – and the message could be unrecognisable.

This kind of jumbling can be done in a complex way, by special, mathematical methods, which would make it impossible to unjumble. In e-commerce, the way the letters have been jumbled, and hence the way they need to be brought back, is based on a long, secret number, which only the sender and the receiver know. The message, hence cannot be opened by anybody but the proper recipient and when the message opens, the receiver would know it came from a specific sender. As any change in the jumbled message would make it impossible to open, the receiver would also know that the message has not been tampered with.

This number that is the basis of the jumbling needs to be very complicated, in fact, a ‘random number’, that is to say, without the digits following a pattern that could be guessed or worked out. Methods like throwing a dice, or the last digit of the number of grains in a handful of sand do generate a highly unpredictable series. These, however, are time consuming and cannot be used for transactions where random numbers are required on the fly! There are hence methods that use a complex, mathematical procedure, to generate an ‘almost random’ number, given a starting input. A limitation of this method is that the procedure generates the same random number every time it gets the same starting input. One example of the input is the time, in milliseconds, when a computer is switched on.

A problem with this particular method is that an intruder could cause the computer to crash, so that it restarts, usually within a minute. Now, the intruder knows that the time of start is one of the millisecond counts within that minute. As a minute has 60 seconds, there are only 60,000 milliseconds to try out, before the start for generating the random number, which is also known as the ‘seed’, is discovered.

The trouble with random numbers which depend on a ‘‘seed’ (these are known as ‘pseudo-random’ numbers), is that they can be worked out once the ‘seed’ is detected. Other sources, often of ‘true’ random numbers, could be the unpredictable fluctuations in the electrical resistance of a conductor, known as ‘thermal noise,’ or even the phase of pulses of light generated by a laser. The methods, however, are cumbersome and have often have limitations.

In contrast, the paper in Nature Communications says, are chemical reactions, where different products of chemical combination, although they arise according to a specific average rate, arise randomly, and could help generate random numbers. The problem, however, is that it is only the average that can be measured – the individual molecules that form cannot be labelled and identified.

It is of this difficulty that the group has succeeded in finding a way around. The DNA molecule, which is there in every living cell, with all the information about the organism, is a clear sequence of specific chemical groups, called ‘side chains’, that are attached to a backbone. In the DNA of a living thing, the groups along the length of the DNA are fixed, and are the same for all the cells of an individual. This is why, when the two strands of the molecule separate, at the time of cell division, the strands are able to recreate the original DNA in daughter cells.

The researchers hence set out not with a DNA molecule, but with a strand of the backbone on which chemical groups attach, to form the DNA molecule. When placed in a bath of the four known chemical groups, the so called, A, T, G and C, whose sequence is what the DNA molecule is, these groups attached to the backbone, one after the other, with equal probability, but quite unpredictably. The sequence of groups in the strand of DNA was hence a true random number – one that arose without the need to be worked out from a ‘seed’. DNA backbones that are allowed to build up the sequence of ‘side chains’ in a bowl of the correct chemical groups, would hence grow into a collection of true random numbers, each one as long as the operative part of the backbone. 'As we now have efficient methods to read out the sequence of groups in DNA molecules, the process could be automated, to create vast libraries.

The group got DNA strands synthesised, in this way, by specialised, commercial laboratories, and the sequences produced were analysed for randomness. After employing well known techniques to eliminate biases that the process introduces, the sequences “passed every test, with a score that surpassed the statistical minimum,” the paper says. As DNA is both compact and robust, and can be preserved and transported with ease, DNA could soon join the forefront of Random Number Generators that are commercially available, the paper says.

------------------------------------------------------------------------------------------ Do respond to : response@simplescience.in -------------------------------------------