Data Genom, Gen AI, and Mie Ayam Daplun
Tauhid Nur Azhar
By leveraging advanced algorithms and machine learning techniques, generative AI is enhancing our understanding of genetic information, leading to breakthroughs in personalized medicine, disease prevention, and biotechnology. Generative AI involves models that can generate new data samples resembling a given dataset. (Nikhil Maske, 2024)
A week ago, I received a WhatsApp message from Dr. Arifa Rahmana Abdulah, SpOG, MHKes, who shared that not long ago, he received an offer from one of the leading mall managers in the beloved city to open a future clinic that not only focuses on health and aesthetics but also regenerative and rejuvenating services, including anti-aging. Isn’t that interesting?
The development of medical service models that may have been new to us recently is certainly present, among other things, due to the rapid progress of medical technology that supports it.
That’s right, because in the next month, I was entrusted with being a motivator for new students in the Biomedical Engineering program at Universitas Telkom, in the form of an introduction to the progress of medical technology in the future, where the core of the development is a synthesis of basic medical science and various innovations that are presented by the progress of research in the biomedical and biotechnology fields.
As a result, I was able to think a lot and try to envision the conditions that will occur in the future, including the map of the road that may be increasingly filled with various surprises that were previously unimaginable.
A few years ago, the trend of predictive medical services had already started to be offered by healthcare service providers such as clinical laboratories.
The rapid progress in genomic technology has made many laboratories acquire biotechnology technology as part of their premium services. The result is the availability of panels of examinations that can be chosen by the public, where the results can describe various disease risks and are followed by analysis results that give suggestions related to changes or adjustments that patients need to make to prevent the occurrence of things that are not expected, or to improve the capacity of physiological potential owned.
One of the leading clinical laboratories in Indonesia, Prodia, since 2019 has offered a Nutrigenomics examination package, where through Prodia Nutrigenomics, more than 50 types of genes and 75 genetic variations (SNP) can be examined, grouped into genes related to nutrition and capacity, as well as compatibility and exercise.
The group of genes related to nutrition encodes several characteristics consisting of taste perception function, response to food, nutrient metabolism, and body weight management and composition.
On the other hand, the examination of the group of genes related to exercise function can provide an overview of the benefits of exercise, strength, and endurance of an individual.
In addition, the nutrigenomics examination panel can play a role as a guide in choosing food for weight loss programs and as a guide for a healthy lifestyle to prevent the occurrence of certain diseases.
One of the genes examined in the nutrigenomics examination is the FTO gene. Where the FTO gene (fat mass and obesity-associated gene) is a gene related to fat mass and obesity.
The FTO gene was first discovered in 2007 in a GWAS study of type 2 diabetes mellitus (T2DM) in Europe. The FTO gene is very important for insulin secretion and beta cell function.
Functional biological studies show that the FTO gene encodes an enzyme 2-oxoglutarate-dependent NAD+ demethylase found in many tissues, but also found in the hypothalamus or the center of energy balance control related to the mechanism of hunger and eating processes in humans.
The FTO gene can affect the amount of fat stored and distributed in the body. Individuals with certain variations of the FTO gene have a 20–30% higher chance of experiencing obesity. Variations of the FTO gene are also related to increased body fat composition and metabolic parameters, as well as metabolic disorders related to obesity such as Diabetes Mellitus (DM) type 2.
Not only that, many other genes can be examined and identified with the advancement of sequencing technologies such as next-generation sequencing and microarray gene chip testing, which is relatively very practical, including if there are mutations or characteristics shown by differences in the level of SNP (single nucleotide polymorphism).
In addition to Prodia, there is also a biotechnology and genetics startup company founded by a leading Indonesian technopreneur, Kiwi Aliwarga, named Widya Genomik, located in the Special Region of Yogyakarta.
This genomic laboratory is equipped with a BSL-2 (biosafety level 2) laboratory facility, similar to what we developed on a large scale during the pandemic a few years ago.
At Widya Genomik, the capacity for examination supported by advanced technologies such as microarray is aimed at producing a screening process against various degenerative risks, including aging and nutrigenomics. Including analyzing DNA methylation and determining polygenic scores to determine the rate of aging, biological age, health risks, suitable nutritional intake, and an individual’s genetic profile, which will be very useful in the context of healthcare services based on a precision medicine approach.
Speaking of genetic testing for polygenic risk assessment for various diseases, it’s undeniable that various technologies and approaches are becoming increasingly precise.
In the complex cardiovascular system with a wide range of spectrums, the role of identifying genetic variations, such as in polygenic testing, is very important and crucial. Because different configurations of genes can cause different pathological conditions of the heart and cardiovascular system. Many genes have causal relationships with various pathological conditions such as cardiomyopathy, arrhythmia, arteriopathy, familial hypercholesterolemia, and Marfan syndrome.
Genes that may be related to diseases in the cardiovascular system include:
- MYH7 gene related to familial hypertrophic cardiomyopathy
- BPC3 SAYA, TNTT2, TNTI3, and LMNA genes related to cardiomyopathy
- Casz1, Rnf38, Tmem161b, and Pde4dip genes related to structural heart defects in the growing and developing phase
- Dnajc18 gene related to heart homeostasis mechanisms
In line with the advancement of genomic testing techniques, some biological and biochemical parameters have also been developed in the previous phase to be used together to determine disease risk factors, such as in cardiovascular diseases. One of the biological indicators is homosistein.
Where a plasma homosistein level of 5 μmol/L is associated with a 60% increased risk of coronary heart disease in men and 80% in women. High homosistein levels can damage the inner layer of arteries and increase the risk of atherosclerotic plaque formation.
Homosistein can also increase the risk of atherosclerotic disease in coronary, cerebral, and peripheral arteries, as well as arterial and venous thromboembolism.
Testing homosistein levels can help doctors from healthcare providers identify vitamin B6, B12, or B9 deficiencies, also known as folate or folic acid. Homosistein testing can be used to help evaluate the risk of cardiovascular diseases such as coronary heart disease and stroke.
So, what is the process of genetic testing using the microarray method? Don’t rush to jump to the technique, a good introductory question is, what is precision medicine that uses the microarray technique?
Precision medicine based on the genome is a medical approach that tailors treatment and healthcare interventions based on an individual’s genetic profile.
One of the technologies used to achieve personalization is genetic testing using microarray. This technology allows the detection of relevant genetic variations to predict health disorders, determine exercise patterns, and choose the most suitable diet for individuals.
Precision medicine based on the genome uses genetic data to make more accurate medical decisions. Specifically, certain genetic patterns can indicate a person’s tendency to develop diseases such as diabetes, hypertension, obesity, and even reactions to specific exercise and diet.
While microarray testing detects small variations in DNA called single nucleotide polymorphisms (SNPs). Where SNPs can be linked to disease risk or specific metabolic responses and play a role in planning various programs and models of healthcare interventions.
If we want to test ourselves to know our genomic profile using the microarray method, our DNA sample is usually taken from saliva or blood. The DNA is then extracted from the sample and amplified using PCR (Polymerase Chain Reaction) technology.
The amplified DNA is then incubated on a microarray chip containing thousands of specific probes for specific SNPs.
After hybridization, the microarray chip is scanned using a scanner that produces quantitative data on each SNP present in the genome.
The data obtained is then analyzed using a specific algorithm to identify relevant genetic patterns related to health risks, exercise, and optimal diet.
Microarray testing can identify an individual’s risk of various chronic conditions such as cancer, heart disease, and metabolic disorders. Testing SNPs related to fat metabolism, glucose-insulin control, and inflammation/inflammatory processes can provide important information about potential medical conditions that allow healthcare professionals to take early intervention actions.
Genetic analysis through microarray technology can also help determine the most effective exercise patterns and the most suitable diet. For example, genes related to energy metabolism, muscle response to aerobic exercise, and metabolic preferences for fat or carbohydrates can be identified. This allows for the determination of whether an individual is better suited to endurance training or strength training.
Currently, the microarray technology used in genomic laboratory services includes the Affymetrix GeneChip platform microarray, which is the most commonly used and widely used, with thousands of SNP probes.
There is also the Illumina Infinium Assay microarray product, which is based on bead array technology that can detect over a million SNPs in a single test.
There is also the Agilent SurePrint Microarrays, which uses the principle of hybridization with high sensitivity to detect SNPs and CNV (Copy Number Variation).
With further development in bioinformatics and multi-omics data integration, microarray can play a very important role in precision medicine. Especially if the analysis results can be integrated with data from other supporting tests, and analyzed with the help of artificial intelligence/AI technology that is increasingly developing various models.
Currently, the darling of the world of AI is the genre of Generative AI, and can the approach of Gen AI be integrated into the analysis of genomic test results, such as those from microarray tests?
Using genomic data, generative AI can help predict the risk of certain diseases based on specific genetic patterns. In addition, Gen AI can analyze genomic data to design therapy that is tailored to an individual’s genetic characteristics.
Through the interpretation of medical images (such as MRI or CT scans) combined with genomic data, Gen AI can help improve the accuracy of diagnosis. Gen AI can also simulate and model the structure of potential new drugs based on genomic analysis.
The generative AI models that can be used in medical and genomic data analysis include, among others, Variational Autoencoders (VAEs), which are used to learn the hidden representation (latent space) of genomic or medical images and generate new data based on the distribution and patterns.
There are also Generative Adversarial Networks (GANs), where GANs consist of two networks, a generator and a discriminator, that compete. This model can be used to generate high-quality medical images or generate realistic genetic sequences.
Then, of course, there are Transformers that can be used in genomic sequence analysis with the ability to understand long-range relationships between nucleotides. Transformers can be applied to DNA or RNA analysis, including, of course, the results of microarray tests.
The systematic process of developing AI-based analytical models or Gen AI begins with data collection using various mechanisms and filtering to ensure that the data is valid, verified, classified, and standardized according to the agreed-upon format.
The data used includes, among others, genomic sequences (DNA or RNA), medical images (MRI, CT scans, etc.), clinical data from electronic medical records (RME), statistical social and economic data from BPS, health insurance data from BPJS, and other relevant data.
Next, the data, such as genomic data, will undergo preprocessing to remove noise, followed by normalization. While for medical images, techniques such as segmentation and filtering are applied to focus analysis on relevant areas.
Then, feature extraction is performed, where the AI model learns the important features of genomic or medical data, as well as supporting data. For example, in genomic data, it can be in the form of relevant SNPs (Single Nucleotide Polymorphisms). This is required for modeling and training.
Gen AI will be trained using processed datasets. Until models like GANs and VAEs can be trained to generate new data similar to the original data or to predict specific results.
After the model is trained, the results are interpreted in the form of visualizations that are easy to understand, such as 3D representations of the structure of new potential drugs, genetic mutation maps, or medical image segmentation for clinical analysis, etc.
The AI model is then evaluated using testing data to ensure the accuracy and validity of the predictions or generated data.
But, to be honest, I wrote the article above because I was inspired by a beautiful surprise I received in front of the big Purwokerto station.
During the muhibah trip to Widya Genomik lab in Sleman Yogyakarta and completing a national task related to the development of a Virtual Poison Center, I chose the Ninja adventurer route by taking on the role of the Bolang train.
A choice that I always enjoy, because not only do I love the world of railroads, but along the way, I can find various Nusantara culinary delights that are so tempting. In addition, traveling by train can be very cheap if the itinerary is right.
For example, yesterday I departed from Bandung using the Baturaden Express train in business class, with very comfortable seats, at a reduced ticket price of Rp. 86,000, to Purwokerto. The transit took about 1.5 hours, then I continued my journey to Jogja using the Jaka Tingkir train.
The journey on the Jaka Tingkir train was enjoyable in the new generation economy class car, refurbished by Balai Yasa Manggarai, which was so beautiful and comfortable. All the happiness on that economy train was priced at Rp. 105,000 to Lempuyangan Station in Jogja.
How is that possible? It’s possible because I used the Go Show scheme that can be accessed through Access by KAI, which offers promotional prices for certain trains that still have empty seats within a 2-hour range before departure time. This scheme is usually applied to long-distance trains that stop at several major stations.
The morning journey with a ticket price under 200,000 (Rp 191,000) with the beauty of Priangan’s natural scenery, the charm of Banyumas with the display of fertility along the Serayu River, the exoticism of historic train infrastructure such as the Notog-Kebasen tunnel, and Ijo, and various legendary bridges such as the Cirahong Bridge in the Tasikmalaya-Ciamis border, and the Serayu Bridge, which now has a cafe on its side that is currently viral, made me feel grateful. What else can we complain about?
But it turns out that the surprise didn’t stop at the comfort of traveling by train. A culinary surprise was waiting to be savored. The 1.5-hour transit in Purwokerto was a gift. With a quick pace like cardio training, I rushed to take photos of the new Kebo Kuning mini locomotive monument in front of the station, withdrew some money from the ATM, and immediately ran to the Mie Ayam Daplun warung on the emper trotoar in front of the Resor Jalan Rel 5.5 Daerah Operasi 5 KAI Purwokerto office. It’s just a stone’s throw away from the main exit.
At the Daplun warung, I quickly ordered a bowl of complete chicken noodle with ceker and a glass of fresh orange juice. When the super delicious chicken noodle was served, I didn’t wait long to add sauce and sambal and eat it with crispy aci crackers.
Instantly, my umami receptors burst and almost exploded. They couldn’t withstand the wave of pleasure that came crashing down with the sensation of 5′-adenosin monophosphate (AMP), 5′-inosine monophosphate, and 5′-guanosine monophosphate (GMP) sodium.
It turned out that Daplun was playing “curang”. It added a special bumbu to its chicken noodle. The special bumbu was a special mixture with a main ingredient of peanuts.
Peanuts, specifically, which contain 42 mg of Calcium, 177 mg of Phosphorus, and 1.4 mg of Iron per 100 grams. The peanuts in the Mie Ayam that are rarely found are what made me want to “mix” and concoct AI in genomic analysis so that the taste is more enjoyable and delicious and unique, and beautiful.
As a result, the article above was written while I was lounging on the guest house’s veranda, the Hadiningrat guest house in Jogjakarta, which is cheap and comfortable, near SMP Negeri 6, Tugu, and Kranggan Market. Of course, accompanied by a cup of Kapal Api coffee, and nasi gudeg for Rp 8,000, and two fried bakwan for Rp 2,000 that were sold around by a mother.
