Biologists really want to lose their jobs... Artificial intelligence has created a fully automated laboratory in Silicon Valley.
Release date: 2017-07-25
The laboratory of Zymergen, a start-up biotechnology company in the San Francisco Bay Area, is unconventional at first glance - the lab's desk is neatly placed in test tubes with various chemicals on the shelves. However, looking around, you will find that the staff in this lab does not seem to be the same, they have some mechanical movements and the sound of fans. That's right, all the robots working in this lab are robots, here is Zymergen's Future Biology Lab.
Here you can see the robot sticking out a robot, grabbing a plastic plate with hundreds of holes on it, and then scanning the bar code on the board. The next step is unobservable by the naked eye. The robot gives a shock wave to the plastic plate, so that each hole can splash a small droplet of liquid. The droplet is quickly taken by the robot and then sent to the next one. The instrument is analyzed. This robot can take 500 samples in one second compared to the traditional use of a hole in a hole in a pipette.
However, Zymergen scientists will tell you that these are not future. The use of robots for experimental operations and recording by barcodes has been widely used. The technique of taking liquid through sound waves, called sonic droplet ejection, has even existed for decades. So where is the “future†of this lab? Dr. Jed Dean, co-founder of Zymergen, will tell you: "I don't know what experiments these robots are doing. Because these experiments are entirely designed by an artificial intelligence program."
That's right, even though robots have a universal application in the lab, it's all about human scientists from hypotheses, design experiments, and analytical data. For Zymergen, their goal is to completely replace the role of human scientists, and these tasks are done by artificial intelligence programs. The development of machine learning in recent times has made it possible for robots to be fully qualified for these tasks.
Zymergen's actual business is to increase the production of microorganisms that can produce important industrial raw materials that can be used in many fields such as biofuels, plastics, and pharmaceuticals. By transforming the genomes of these microbes, Zymergen wanted to find a number of ways to increase yields.
However, these industrial microorganisms have been selected and optimized for many years. In order to further increase the yield, it is necessary to conduct a very in-depth study of their genomes, and further research based on preliminary data. This is a time-consuming and laborious task for even very good scientists. Zymergen CEO Joshua Hoffman estimates that a scientist can complete about 10 experiments a month. With the help of traditional robots, this number can be increased to 1,000 items per week, provided that the correct experimental design and instructions need to be provided to the robot, which is often a bottleneck.
For a microbe with 5,000 different genes, if there are 10 variants per gene, then at least 50,000 different microorganisms need to be tested. You can find dozens of which can increase your yield, and then test combinations of these different individual gene variants. Even if you can test 1000 kinds per week, this is still a long process. In addition, the light test yield is not enough, and the growth capacity of the microorganism itself needs to be closely watched. And this is where machine learning can make a difference. From the results of the previous experiment, the machine learning algorithm can design the hypothesis that the next step needs to be tested.
Mr. Hoffman said that Zymergen has been able to increase some microbial production by 10% so far. This seems to be a small number, but for the $160 billion chemical industry, a 10% increase could mean billions of dollars in economics.
However, artificial intelligence can find ways to increase production, but it does not know the biological mechanisms behind it. Dr. Dean believes that this is actually one of the advantages of artificial intelligence. Because traditional scientific methods only start with known genes associated with the production of specific products, this has significant limitations. Zymergen's experiments found that many of the genetic changes found by artificial intelligence to increase yield are not directly related to their chemical synthesis pathways, and even many of them have unknown functions.
Many aspects of science can be done by artificial intelligence (Source: Science)
For biologists, these discoveries are exciting. Because they can start by reverse the reasoning by increasing the yield, which is expected to uncover more of the functions of genes that are not known. There is also a more ultimate possibility, that is, one day artificial intelligence programs can find the mechanism behind it. By then, I am afraid that human scientists really have no use.
Reference material
[1] A new breed of scientist, with brains of silicon
[2] Zymergen official website
Source: Health New Vision (Micro Signal HealthHorizon)
Natural pigments are food pigments obtained from natural resources. Pigments extracted mainly from animal and plant tissues and microorganisms (cultures), in which vegetative colorants predominate. Natural pigments not only have the function of coloring food, but also have physiological activity.
Natural pigments generally come from natural ingredients, such as beet red, grape and pepper. These foods have been recognized and accepted by the majority of consumers. Therefore, natural pigments from these food sources are more favored by consumers and safer to use. Most of it comes from plant pigments.
Our company can provide caramel, carmine orange, gardenia yellow, red koji, chili red, turmeric, β -carotene and other natural pigment products.
Natural Pigments,Natural Color Pigment Capsanthin,Food Pigment Capsanthin,Natural Pigment Gardenia Yellow
Allied Extracts Solutions , https://www.alliedbiosolutions.com