Genomics and role of Artificial Intelligence for Better Healthcare
According to a paper by Ron Sender et. al in 2016 estimated human body is composed of 3.0*1013 human cells. These cells consist of many components, but we are interested in DNA here, the majority of DNA lies inside the nucleus and some amount in the mitochondria of the cell. And all the genes contained inside a cell specifically DNA is referred to as a “genome”.It is the machine level code of the human body. A genome consists of all the information required to build, function, and maintain that organism. These DNA consist of long paired strands of 4 base i.e, Adenine, Thymine, Guanine, Cytosine which are the code that instructs cells specifically RNA, protein, and body how to behave. Hence they are the code of the human body. The human genome is 99.9% the same in all people but has a variation of 0.001% in humans. But closely related peoples can have more similarities. So this means I am 99.9% similar to Albert Einstein or John McCarty, but that 0.001% difference contributes to differences in health, appearance, behavior, thinking, and many more.
The A, T, G, C DNA base forms a large sequence where A can only be linked with T and G only with C or vise versa. Genomic Sequencing is the decoding of sequences of A, T, G, C bases from the human DNA and genomics is the study of genomes and their interaction with each other as well as with the environment. For example, a DNA sequence can be in the form like this where A can only be paired with T and C only with G.
A human body has approximately 3 billion base pairs which are equivalent to approximately 715 MB of data. How much it would look like during visualizing you can imagine. Due to the massive benefits from genome sequencing approximately 3.8 billion dollars were spent in the “Human Genome Project” which involved hard work of global scientific community collaboration from 1990 to 2003.
But the recent advancement of Artificial Intelligence specifically its sub-branches machine learning and deep learning made it simpler to achieve and additionally unlocked more additional benefits.
The massive data of human genomes 6 billion letters sequence cannot be studied by human eyes to extract relevant knowledge from it but can be done rapidly and easily by artificially intelligent machines. If you want to know how there are machine learning and deep learning approaches for it.
As any machine learning model uses data to train it to identify the pattern in genomic sequence and based on the pattern and genomic sequence prediction, an estimation can be made. These potentials of Artificial intelligence in Genomics have opened following new possibilities.
1- Produce a Genomic Sequence Rapidly and Affordably:
Before the existence of AI and lack of knowledge about the potential of AI it took researchers 13 years to sequence a genome. But now it can be done at a cost of less than $1000 dollars as per the National Human Genome Research Institute. Highly throughput sequencing also known as Second Generation Sequencing in 2000 made DNA Sequencing more rapid and DeepVarient by Google improvised accuracy using the same technique.
2- Genomic Sequencing to Predict Diseases or Future Complications
What disease a person will have in the future or whether a person has any specific disease can be predicted accurately by analyzing the genomes and AI. There are different approaches but widely used are (i)Interpreting Variance: In it, the variance in the genomic sequence of the patient being tested for the disease is compared with the genomic sequence of a normal healthy person and genomic sequence of the person with the specific disease using Artificial Intelligent Techniques specifically deep learning models, these trained deep learning model look for patterns, and based on obtained patterns the model will predict the possibility of developing a particular disease.
3- Precision and Personalized Diagnosis, Treatment/Medicine:
Genomics allows us to understand the person’s body in larger depth, understand mutations, visualize, and recognize patterns in sequence for diagnosis. The pattern in the genomics sequence of a patient is compared with the genomic sequence of diseased and normal humans.
Let’s understand the illustration above about personalized diagnosis, as you can see figure 3 shows the genomic sequence of a completely normal person, but figure 1 and figure 2 shows that genomic sequence is varying at some specific locations in the series in case of the cancer patient and diabetic patients. Similarly, the above technique is used to diagnose a person for a specific disease. Now talking about personalized medicine means recommending medicine to a patient by properly studying his/her genomics, medical history to suggest the best medicine for his/her better cure. This involves analyzing a massive amount of data by an artificially intelligent machine to predict that this specific person with these genomes will respond positively to medicine A and another person with another type of genomic sequence will respond better with medicine B.
In the future, artificial intelligence will enable highly accurate predictions about potential health complications. A newly born child “kundali” about his/her health can be made at birth. Gene editing will be possible to cure diseases. Genomics backed by artificial intelligence is the future for preventive, diagnostic, and curative medicine. Human body machine level codes i.e, Genomes can be debugged in the future to remove the bugs i.e, diseases in the human body.
