Computer super-meta-analysis
Silicon Scientist seeks Cancer Cure
They were allowed to the control panel later... The reality is a little different: nurse robots are rare, only 159 units were sold in 2012 on the whole planet, but automation controls power systems, drives airliners, independently forms data center structures. And now "software robots" are beginning to take on the work of a scientist, the highest kind of intellectual activity!
Medical research has always been the most relevant in human society. Both from a moral and ethical point of view, and from a pragmatic and financial point of view (in healthy states, a much larger share of GDP is allocated to healthcare than to defense and the police). Well, cancer research occupies a very special place among them. We are made up of von Neumann machines, which biologists call cells. And violations of their reproduction processes, which belong to the group of oncological diseases, are terrible, and their statistics are terrible. It is not for nothing that the "cure for cancer" occupied a firm place in dreams and forecasts…
And now "smart" machines are being included in the fight against the terrible disease. Moreover, they are included at the stage of scientific research, that is, exactly where discoveries can be made and technologies that can save millions of lives can be developed. We are talking about software created jointly by IBM and Baylor College of Medicine (Baylor College of Medicine). This educational institution in Houston (Texas) specializes in the most high-tech types of medical research and traditionally pays great attention to computer technology, having considerable capacities.
And the new software is engaged in a very interesting business. It "reads" scientific articles, in the amount of tens of thousands of pieces. And – "digs out" new scientific knowledge in them! Pilot operation was carried out on the subject of protein p53. This protein was christened by its molecular weight – 53 kilodaltons. And its function in the body is extremely important. It works as a suppressor, suppressor of oncological processors. It is he who "guards" the body from malignant tumors! And in about half of the cases of this disease, violations of its production are noted due to mutations of the TP53 gene.
So the p53 protein was depicted in the company of DNA.Well, at the level of the biochemistry of the body, the behavior of the p53 protein is regulated by enzymes called kinases.
And these enzymes are usually targeted in the treatment of cancer: adjust them – and the p53 protein itself will begin to do its job. Let's give an analogy with another large system – the state. Changing the work of enzymes is equivalent to establishing correct and enforceable laws, under which the activity of people – always self–serving and not inclined to foresee the consequences - begins to work for the good.
And with such funny pictures, biologists illustrate the role of p53 in the body.Well, what prevents you from "correcting the laws" in the body?
So exactly what it is is a great system. There are too many processes, on too many levels, with too many too complex connections. Moreover, the technology has provided methods for tracking these processes. And there are a lot of biochemical studies being conducted… It is only necessary to bring these data into the system, to extract from them knowledge useful for medical research and – at the next stage – for practical medicine.
And that's where the "smart" car comes to the rescue. How many scientific articles can you read in a day? Not to run diagonally, but to read? Three, five… Maybe ten? And there are two and a half hundred working days a year… And the volume of work being carried out in the world is enormous. What part of what others have already done will you master? You can, of course, follow the path of specialization… But the trouble is: nature has no specializations. Well, at the level of zoology, the division into the Zahoderovsky "Academician on whales, / Academician on cats" will work: a whale does not even fight an elephant in practice, although who knows from whom cat food is made…
But at the level of the biochemistry of the body, the symphony of proteins and enzymes continuously playing in it, the number will not pass. Focus on one or another protein, gene, enzyme – and you will find yourself in the role of Pushkin's Salieri. But music is like a corpse can be torn apart: it doesn't hurt music, and it doesn't have self-awareness. And then the body… Who really doesn't want to become a corpse. And so technology came to the rescue – just as it once gave the microscope to the natural sciences, giving rise to microbiology with its most useful practical conclusions.
Now the software from IBM and Baylor Medical College is starting to work as an amplifier of cognitive abilities. The results of the work are impressive. Sixty thousand articles on biochemistry have been read. All mentions of known kinases are highlighted – those enzymes that interact with the p53 protein. Then, probably (details of the work are given with accuracy: this is not just science, but also engineering with business), some common features of these enzymes have been identified. Well, on the basis of common features, a list of enzymes of interest to this problem was formed.
Well, then the predictions of the computer researcher were tested in practice. And it turned out that he pointed out quite promising areas of experimental scientific work. Anyway, according to Oliver Lichtarge from Baylor College of Medicine, seven out of ten enzymes predicted by the robot scientist are very likely to be kinases. Of course, the findings still have to go through a detailed study and a cycle of publications in peer-reviewed scientific journals, then the increase in the pace of research is impressive.
The fact is that on average in the twenty-first century, according to the same Lichtarge, one kinase is opened per year. And here – seven at once! The acceleration is amazing. After all, these are potential medicines. This is an opportunity to save human lives. And this is a way to reduce the cost of developing new drugs, which now ranges from half a billion to a billion dollars, in no small part due to the fact that the work is being carried out blindly. Because of the too great complexity of our organisms, and because of the huge amount of data accumulated by researchers.
Everyone has heard the saying "Life is short, art is eternal". As a rule – at the wake of an artist. Although in Latin it – "Vita brevis, ars longa" – means "Life is short, art is vast"… But even this is inaccurate, the words were said by a great doctor. And Hippocrates said: "Life is fast, and [medical] science is huge." So it was in antiquity, so it remains in modern times. Education should last a lifetime, and you will learn only a small part of what you should…
And today, according to Lawrence Hunter, head of the Center for Computational Pharmacology at the University of Colorado, author of the most interesting book "Artificial Intelligence and Molecular Biology", over the past year, more than a million biomedical works have been added to the US National Library of Medicine, according to the Medline database. And in total there are more than 23 million works. The ratio between brevis and longa is not at all in favor of a person…
But, as we said above, information technology comes to the rescue. They help to extract those facts that otherwise would have been – together with the researchers who found them with difficulty – forever buried in piles of papers and chains of bytes on bookshelves or servers... Of course, these are only outlines for future experimental work by classical methods of natural science research, in this case biomedical. But the increase in the rate of extraction of useful knowledge seems to be very significant!
And, of course, these successes of IBM and Baylor have, as always, quite obvious predecessors in science and technology. First of all, this is the technology of data mining – the excavation of data that Computerra has been writing about since the mid-nineties. Next – the scientific "model of the world". If in the Copernican model the Earth turned around the Sun, and Kepler's planets ran along ellipses, then the articles processed by the "smart machine" contained a model of biochemical interaction in the body described in a mixture of natural and scientific languages. And the "run" of this model gave new knowledge!
And this particular story, which we have dwelt on in detail, means, in addition to all that has been said above, the beginning of a new branch of business. Cognitive entrepreneurship – is that how to translate "The Cognitive Enterprise"? A business based on the development of knowledge – its possible scale can be estimated by the above billion-dollar costs for the development of a single drug!!! And this is already a specific area of application of forces and money, conferences like those held on November 19, 2013 by the IBM Research – Almaden laboratory in San Jose are already taking place on it. A new branch of the IT business is maturing before our eyes!
Portal "Eternal youth" http://vechnayamolodost.ru28.11.2013