The Symbiotic Relationship Between AI and Biosciences
While I support the motion stated above as the title. I am here to convince and not confuse you about this motion.
Outline:
- Biosciences
- AI — A brief introduction to Artificial intelligence, and its different applications across various fields with relative examples.
- The interdisciplinary relationship — between AI and Bioscience
- Conclusion
Bioscience
Bioscience (the science of life) consists of various fields, as well as wide-ranging. For example, biochemistry is a combination of biology and chemistry. Biotechnology mostly deals with the application of technology and biochemistry. Biomedical science majors in scientific research, it can also be defined as a bond between Biology and Medicine, because of this, bioscience can cut across health science with the involvement of medical science. For this reason, it will be unreasonable to ignore health science in this discussion.
Biomedical engineering is another sector within the bioscience field that deals with the application of engineering techniques and principles to biology and medicine. All of these fields are somewhat intertwined with just a fraction of the same continuum
AI
When a child is given birth to, the brain is not fully programmed yet, until it begins to interact with its environment, then the brain begins to learn from this interaction.
Before the generation of computational techniques which has brought about the blessing of Artificial Intelligence (AI), human intelligence was the yardstick(it still is).
AI was inspired by human intelligence, and as every definition will explain, AI is basically a mimic of the human brain, which means before the success of AI, the human brain had been thoroughly examined. In the achievement of AI, cognitive neuroscience (sub-field in Neuroscience) — a science that studies how the human brain works — was of great influence.
We’re well aware of the progression of AI all around us, from the tailored trends on our social media platforms like Twitter, recommended movies on Netflix, the traffic information on Google Maps, the filtering of our email (spam detection), even recommendations on our favorite online stores. We should also be aware of devices like the IWatch that take records of our daily routine to notice abnormal trends in our heart rate or calories intake.
From the various applications of AI stated above, we should notice how AI has cut across majorly all disciplines, from finance to social activities, agriculture, and even medicine. The Bioscience field is not an exception, from the creation of drugs to the detection and identification of certain diseases, AI has also influenced a significant growth in biological and life sciences.
The Interdisciplinary relationship between AI and Bioscience
The truth is, like mutualism, AI and Bioscience benefit from each other. But, the relationship between them is more commensal because bioscience benefits a lot more from AI than AI will ever benefit from bioscience.
Asides from neuroscience which influences neural networks (Deep Learning), a significant part of AI, and that's about it. AI/ML on the other end has further helped in research made in bioscience, medical science, and a lot more. This explains the “symbiosis” in their relationship. Commensalism!
In different fields of bioscience; from biochemistry to microbiology, biotechnology, and so on, these data get too big for the scientists to keep up with, research needs to be more productive, and encouraged to run at a fast pace, results, at a faster and more productive pace, yet, the data keeps getting bigger. That is surreal if you ask me! But nobody did:).
Artificial intelligence disagrees with this, and with AI, algorithms were built depending on the problem the software is solving. This compels AI/ML to serve fundamentally in different fields of Biosciences, from progress in research to drug creation through medical science, and principally push the evolution of scientific discovery through the availability of technology, making life better for the masses on a wider prospect, enhancing productivity and speed, also through computational biology.
Here are examples of the application of AI in Bioscience:
1. Drugs Creation:
I will give a typical example of how AI has proved useful in drug creation. Over decades, science has tried to identify different structures of protein due to protein folding. But out of over a 200million proteins, biologists could only identify 170,000 proteins (confirm here) because not only does it require a great number of decades of hard work, but multi-million dollar equipment is also required. Because these structures cannot even be identified, it becomes impossible to know how to solve the problem they cause in the body.
You are lost? Okay! In a layman’s understanding, Protein is basically what makes everything in the human bodywork, they are complex molecules that make up every part of every organism, from humans to viruses and so on, protein gives instruction to produce DNA.
Protein isn’t limited to that nutrient you get from cheese or egg, that’s not where it stops. The repair of worn-out tissues, formation of white blood cells (Immune Cells), red blood cells (Hemoglobin) is all the roles of protein in the human body, as well as transportation of food nutrient within the body, formation of hormone (ADH, insulin, glucagon), it is why you are employed to take more protein for quick recovery after suffering an injury, and why your grandparents are advised to be fed protein more to promote a healthier lifestyle.
You see how big of a deal protein is?
Furthermore, like letters A, B, C make up the English alphabets, alphabets make up words, words make up sentences, amino acids make up protein such that the perfect combination of amino acids sequence, and the correct structures which these proteins fold is going to define the function of that protein. But, incorrect folding of these proteins results in different kinds of allergies in humans as well as diseases like Parkinson’s disease, Alzheimer’s disease, Cystic fibrosis(CF), and many more.
You remember how big of a deal protein is? So, since proteins always bind together, proteins made from cancerous cells would also bind with free cells and since the body cannot differentiate between the cells, or stop the binding process, it keeps binding and spreading the tumor in that part of the body. Since their different structure cannot be identified, it doesn’t make it any easier to create a solution to these problems.
You have no idea which protein is causing a particular infection or disease, even though scientists discovered the actual structures of some proteins with techniques such as nuclear magnetic resonance and x-ray crystallography, the data is barely enough to make inferences.
No solution can be created based on hypotheses in science. Moreover, why settle for hypotheses when computers and algorithms can give us the idea without us investing so much in experiments? I’m sure the next question on your mind is how AI has helped solve these problems.
Alphafold, an AI system formulated by Google's Deepmind helped identify and discover different structures of protein through deep learning neural network methods, not only in a reasonable amount of time (from a few minutes to a few hours) depending on the different structures of different proteins, and with no multi-million dollar equipment just a deep learning neural network.
You know the way a neural network is trained to identify a certain subject in a provided image through object identification, by providing the network with the appropriate training dataset to understand the subject such that when given test data, it correctly identifies the subject.
Our deep learning model was given the identified protein structures as training data and with a very low bias, the model predicted the accurate value when the test data was input, that is, the model’s prediction of the new protein structure (test data), when compared to the experiment carried out by the scientists, provided almost 100% accuracy.
Identification of these proteins will enable biologists to understand some diseases (the cause, origin, solution…) better and hasten the production of effective drugs to cure or even prevent them.
Imagine how much more AI will help the discovery of new drugs, enhance effective research and play significant roles in creating a cure for diseases that are classified as terminal right now.
In addition, because AI fully understands the data of a patient, it is able to help with the production of the perfect drug, alongside biologists. There are also more ways AI/ML will prove significant in the discovery of new drugs as time goes on.
• Early identification and diagnosis of genetic diseases such as; cancer, diabetes
Identifying cancer, a terminal disease the human race has battled for long did not use to be detectable at an early stage. Pathologists can now detect and diagnose colorectal cancer accurately by analyzing tissue, and also identifying patterns in breast cancer early, with the help of AI’s Artificial Neural Network (ANN).
An accurate analysis, diagnosis, and detection of cancer cells and their pattern give oncologists useful insights to work on the adequate treatment of cancer with the help of ML and deep learning, because of the enormous data on the different types of cancer, and other genetic diseases at large, in addition to the facilitation of faster treatment.
ML coupled with Computer Vision could give radiologists a good view/knowledge of the progression of the patient’s tumor, a good application of virtual screening.
The early detection of these diseases in patients is progress for the human race, it gives the patients a better shot at living and chemotherapy. Previously, immunotherapy has been our most significant approach to fighting cancer, the body’s immune system is tapped to fight cancer.
However, Someday soon, very soon, with AI, scientists would be able to program cancer-fighting cells, but for today, researchers are still at work and more revelations and innovations are set to be made.
In the treatment of genetic diseases such as sickle cell anemia, cystic fibrosis, HIV, and even cancer, scientists innovated CRISPR/Cas9, a gene-editing tool. Because these diseases occur as a result of a mutation in a patient’s DNA, this technique joins the Cas9 with the Guide RNA to locate the infected DNA, either to delete or correct this infected DNA. CRISPR/Cas 9 can also solve the problem of incorrect folding of a protein.
With more research work put into this technique, it only gets better, understanding the side effects, and other significant processes. Another good thing about the support of AI is that the proper algorithm can be used to predict the output of any result of these treatments at any point.
• Research:
This is one area where life science and AI find a certain balance. Biochemistry, or biomedical science, for example, involves research work, the discovery of new data, and new ways to solve a new or an old problem. However, with AI, and a good knowledge of our problem and data, answering questions asked by science with the help of science eases things.
Practically all that has been stated above is as a result of the research work of scientists in combination with AI. Unlike humans, AI software does not have emotions, they are more logical in the process of research, and as a result, there is a very low bias in the output value. AI is also applied in scrutinizing scientific literature.
Recently, ML models have been trained in ophthalmology to discover symptoms of diabetic retinopathy which is the fastest-growing cause of blindness around the world. This is an application of AI's retinal imaging.
Regardless, there are more scientific innovations and progress to come in these interdisciplinary fields. Scientists are researching the ability of plants to efficiently absorb sunlight for its use. If there is a positive outcome in the result of this research, and it spells success, humans might be able to translate that into the creation of synthetic DNA with better solar cells.
On a final note;
The whole point of this article is not to make you see reasons why a certain field is more useful than the other or how one field can survive independently, but to understand that the combination of both fields can yield an extraordinary output.
You know the saying two heads are better than one? It is very applicable here. The application of AI in solving problems in biological sciences is a practical example.
Biology provided the foundation for AI to explore the historical data in the bioscience field. With the application of AI, Machine learning, Computer vision, or any other subsets of AI, this data is used to make better predictions.
In conclusion, since commensalism has always been about one organism benefitting more, and I likened the relationship between AI and bioscience to commensalism. Therefore, Bioscience benefits more in this relationship.
I hope I have been able to convince and not confuse you about the symbiotic relationship between AI and Bioscience.
References
- https://singularityhub.com/2021/11/23/how-ai-is-deepening-our-understanding-of-the-brain/
- https://www.healthline.com/nutrition/functions-of-protein#TOC_TITLE_HDR_2
- https://www.nature.com/articles/d41586-020-03348-4
- https://www.genome.gov/about-genomics/educational-resources/fact-sheets/artificial-intelligence-machine-learning-and-genomics
- https://www.genengnews.com/insights/trends-for-2020/artificial-intelligence-is-helping-biotech-get-real/#:~:text=AI%20applications%20in%20biotech%20include,and%20manage%20clinical%20trial%20data
