GULTIK Prof HR

Tauhid Nur Azhar

Photo by Syauqy Ayyash on Unsplash

Jensen Huang, the Rp 2.005 trillion man, founder of NVIDIA, the producer of GPUs now essential for developing AI technology, enjoyed the legendary gulai tikungan (gultik) dish at Blok M, South Jakarta. He was accompanied by Indosat Ooredoo Hutchison (IOH) CEO Vikram Sinha and journalist Najwa Shihab. Jensen’s expression was enthusiastic as he tasted the gultik. “This is amazing,” Huang said with wide-eyed wonder.

Najwa, Vikram, and Jensen sat on simple chairs and tables, just like the general public enjoying gultik at Blok M. (CNBC, 17/11/2024, 13:30).

After reading this news snippet following an 8 km bike ride around Bandung in search of a legendary chicken noodle dish hidden in a secluded location, I fell asleep under a Mahogany tree and dreamed.

In my deep sleep under a tree that produces abundant oxygen (research from the US Forest Service shows that a healthy adult tree can produce between 260 and 600 liters of oxygen per day), I, exhausted, dreamed beautifully. I dreamed of seeing Jensen Huang, Elon Musk, Sam Altman, and Kang Muhammad Yusuf from MIPA Unpad, sitting around Prof Hammam Riza, the Chairman of KORIKA. In the same location, gultik Blok M. Only this time, Najwa and Vikram were not present.

Apparently, Prof Hammam successfully invited his friend, the LLM-based Transformer expert who gave birth to the revolutionary cognitive aid tool GPT, to gather and discuss with Elon Musk, the founder of Starlink and Neuralink (brain chip), and Kang Muhammad Yusuf, the molecular modeling expert, to sit together and solve one of the world’s major problems: the food crisis.

A crisis also triggered and exacerbated by environmental degradation due to excessive land exploitation, leaving the Earth increasingly powerless due to the loss of resource potential. Looking at the data, Prof Hammam felt that human civilization was at the end of its era and on the verge of collapse.

Thus, at the Blok M corner, an open meeting (literally) was hastily conceived due to the increasingly unclear world situation. The four great sages are the true saviors of the world. In my dream, Prof Hammam was like a Marvel character, an agent of SHIELD saving the world: Nick Fury. Prof Hammam and his KORIKA were Indonesia’s SHIELD.

If we use an acronym, then at Najwa’s gultik gathering, there were: Prof HR, JH, SA, EM, and MY. Four figures who were very systematically considered with great care by Prof HR and his team of AI experts and practitioners from Indonesia.

Gultik was chosen not because of its taste or the alluring location but because through gultik, Prof HR would explain his brilliant concept for saving the world.

In short, Prof HR gave a brief introduction by referring to global statistical data on demographics, food production, epidemiology, and geophysical meteorology.

Then Prof HR went straight to the point by using gultik as his presentation medium. Prof HR explained the nutritional content and elementary components contained within it, including the active compounds found in a serving of gultik and a cup of coffee, tea, or chocolate chosen by each of our superheroes.

Prof HR explained that in one serving of gultik, there are spices including lemongrass, nutmeg, ginger, red and white onions, and cinnamon. Lemongrass contains active compounds such as essential oils, alkaloids, saponins, tannins, flavonoids, anthraquinones, and polyphenols.

The essential oil from lemongrass (Cymbopogon nardus L) contains compounds such as citronellol (30–45%), geraniol (55–65%), geraniol acetate (3–8%), citronellol acetate (2–4%), L-Limonene (2–5%), terpene alcohol (2–5%), cadinene (2–5%).

Citronellal and geraniol compounds in lemongrass oil have repellent properties that can repel insects. The citral compound in lemongrass stalks has anti-cancer properties. Meanwhile, antioxidants in lemongrass, such as chlorogenic acid and isoorientin (Isoorientin is a flavone C-glycoside with the molecular formula C21H20O11 and molecular weight of 448.4 g/mol. It is also known as homoorientin, luteolin-6-C-glucoside, and 4261–42–1.), can help prevent cellular dysfunction in the heart’s blood vessels.

Another example is the nutmeg seed, which contains active compounds such as Linalool (Beta-linalool, linalyl alcohol, linaloyl oxide, allo-ocimenol, and 3,7-dimethyl-1,6-octadien-3-ol), essential compounds that can help regulate the cardiovascular system, myristicin, an organic compound that can affect the neuropsychiatric system. Then there are flavonoids, tannins, eugenol, and isoeugenol, compounds that act as antioxidants.

Its essential oil contains monoterpene compounds such as Alpha Phellandrene, Beta-Ocimene, Sabinene, 2-beta-pinene, Myrcene, beta- Phellandrene, Trans-Beta Ocime, Alpha Terpinene, p-cymene, Limonene, Gamma-Terpinene, Alpha- Terpinolene, Terpinene-4-ol, and Myristicin. Additionally, nutmeg also contains oleoresin, which is widely used in the food, cosmetics, and pharmaceutical industries.

Meanwhile, its fruit (Myristica fragrans Houtt) is a spice plant with various active compounds, such as antimicrobial, antibacterial, antioxidant, antifungal, and anti-inflammatory properties, and significantly, its sedative effect.

From the drink chosen by SA, hot chocolate, for example, there are active compounds such as Theobromine, which has effects similar to caffeine, increasing energy, alertness, and improving mood. Theobromine is a chemical from the alkaloid group found in cocoa plants. Then there is Phenylethylamine (PEA), which increases endorphin production in the brain, known as the happiness hormone, and Serotonin, a natural chemical that helps reduce stress and anxiety.

Additionally, chocolate also contains vitamins B5, B2, B6, B1, and B9, as well as minerals such as magnesium, copper, iron, manganese, and zinc. Complete, right?

In essence, according to Prof HR, food, whatever its variety and type, functionally is:

1. A source of nutrients derived from its constituent components.
2. A source of delicious taste, savory, umami, sweet, salty, etc., which motivates people to consume it.
3. A source of vital energy.

If point 1 is related to the metabolic process in the realm of anabolism, then point 3 is related to catabolism and the production of vital energy from ATP molecules through the glycolysis, Krebs cycle, and oxidative phosphorylation pathways.

Point no. 2 is related to the taste produced from the interaction of taste molecules with taste receptors in the human nervous system. Taste molecules are chemical compounds that influence the sensory perception of food, interacting with receptors on the tongue and producing experiences of sweet, sour, salty, bitter, and umami tastes. The identification of taste molecules is not only important in food development but also in pharmacology, nutrition, and health. But it seems Prof HR has another meaning in this context.

Taste is one of the main aspects determining human food preferences. Taste molecules, such as sugar (sweet), sodium ions (salty), organic acids (sour), alkaloids (bitter), and glutamate amino acids (umami), play a crucial role in this mechanism. Advances in molecular analysis technology enable the identification and characterization of taste molecules with high precision.

Taste perception is based on the interaction of taste molecules with receptors on the tongue, as outlined in the following theories:

1. Molecular Binding Theory Taste molecules bind to specific receptors (e.g., T1R receptors for sweet and T2R receptors for bitter) located on the taste buds of the tongue.
2. Taste Transduction Theory The binding of taste molecules triggers electrical signals through transduction pathways, such as the cAMP pathway for sweet taste or the PLCβ2 pathway for bitter and umami tastes.
3. Molecular Synergy Theory Some taste molecules work synergistically to produce complex taste experiences, such as the combination of acid and umami in fermented foods.

In a scientific context, there are several methodologies that can be used to map taste molecules and energy and anabolic molecules. These methods include High Performance Liquid Chromatography (HPLC), which can separate molecules based on their interaction with stationary and mobile phases. It can be used to identify sugars, amino acids, or peptides related to sweet or umami tastes. An example of its application is the analysis of glutamic acid in fermented foods.

Then there is Mass Spectrometry (MS), which identifies molecules based on their molecular mass. Its uses include characterizing volatile compounds that provide aroma and contribute to the taste of food. An example is the identification of bitter alkaloids in coffee.

Next, there is Gas Chromatography (GC), which can separate volatile molecules based on their volatility. This method can be used to analyze aroma and taste molecules related to volatile compounds. An example is the determination of ester compounds in tropical fruits.

Following that, there is Nuclear Magnetic Resonance (NMR), which can identify molecular structures based on the magnetic properties of atomic nuclei. It can be used to identify complex components such as polyphenols. An example is the analysis of tannins in tea and wine.

There is also the Bioassay Taste, a biological assay to detect the interaction of molecules with taste receptors. It can be used for screening new taste compounds using cellular or animal models. An example is the evaluation of artificial sweeteners.

And the most advanced is Omics Technology such as Metabolomics, which can be used for profiling small molecules related to taste in food. Then there is Genomics and Proteomics for studying genes and protein receptors for taste.

How can the taste molecules in various foods and beverages be perceived by the human brain? They can even be preserved as memories that help determine preferences in choosing and processing food, even becoming a reward in the motivation process to obtain a certain culinary experience related to taste and happiness.

Taste is the result of a complex interaction between taste molecules, receptors on the tongue, and sensory centers in the brain. This mechanism begins with the introduction of taste molecules to the interpretation of sensations by the brain, involving a series of biophysical and neurophysiological processes.

1. Interaction of Taste Molecules with Gustatory Receptors

a. Specific Taste Receptors: Taste is located on the membrane of sensory cells in the taste buds (taste buds) on the tongue. There are five main categories of taste recognized by specific receptors:

Sweet: Activated by sugars (glucose, fructose) and artificial sweeteners. Recognized by T1R2-T1R3 receptors.

Salty: Activated by sodium ions (Na⁺) through the ENaC (Epithelial Sodium Channel) ion channel.

Sour: Triggered by hydrogen ions (H⁺) that activate the PKD2L1 ion channel.

Bitter: Responded to by T2R receptors, which are sensitive to alkaloids.

Umami: Triggered by glutamic acid through T1R1-T1R3 receptors.

b. Molecular Binding Mechanism: When taste molecules interact with specific receptors, there is a conformational change in the receptor protein.

In sweet receptors, the binding of molecules such as sucrose triggers the activation of the adenylate cyclase pathway, increasing cAMP levels within the cell.

In bitter receptors, alkaloid molecules activate the phospholipase C (PLCβ2) pathway, producing inositol trisphosphate (IP3), triggering the release of calcium ions (Ca²⁺) from the endoplasmic reticulum.

1. Signal Transduction in Sensory Cells

Chemical signals received by the receptors are translated into electrical signals through transduction:

Released Ca²⁺ ions increase membrane depolarization.

This depolarization triggers the opening of additional ion channels, such as TRPM5, which amplifies the signal. An action potential is formed and transmitted through the axon of the sensory cell to the afferent nerve.

1. Neurophysiological Pathway to the Brain

Signals from the gustatory receptors are transmitted through three main nerves:

1. Facial Nerve (Nervus VII): Carries signals from the anterior two-thirds of the tongue.
2. Glossopharyngeal Nerve (Nervus IX): Carries signals from the posterior one-third of the tongue.
3. Vagus Nerve (Nervus X): Carries signals from the epiglottis and pharynx.

These three cranial nerves synapse at the Nucleus Tractus Solitarius (NTS) in the brainstem, which serves as the initial processing center for taste signals.

1. Processing in the Brain

After being processed in the NTS, the signals are relayed to the following brain areas:

a. Thalamus (Nucleus VPM) The thalamus acts as the main relay, directing taste information to the sensory cortex.

b. Insula and Operculum of the Gustatory Cortex This area is located in the insular lobe and is the main center for taste perception. Activation here allows us to recognize different tastes, such as sweet or bitter.

c. Limbic System Components of the limbic system, such as the amygdala and hippocampus, connect taste with emotions and memory. For example, a sweet taste might evoke positive memories.

d. Orbitofrontal Cortex (OFC) The OFC integrates taste, aroma, and texture information, creating a holistic eating experience.

1. Molecular Mechanisms at the Brain Level

a. Neurotransmitters Electrical signals from sensory cells are converted into chemical signals through the release of neurotransmitters, such as:

ATP: The primary neurotransmitter in taste signal transduction.

Serotonin and GABA: Play a role in modulating taste signals in the central nervous system.

b. Activation of Neural Pathways

In the gustatory cortex, molecular signals are translated into specific neuron activation patterns, creating a particular taste perception. These signal pathways also interact with dopamine receptors in the OFC, determining the reward or satisfaction aspect of taste.

According to Prof HR’s hypothesis, ultimately, all processes from taste receptors in sensory organs to the brain are a matter of bioelectric configuration. Where the bioelectric profile that has been well-identified for each taste stimulus and its intensity can be replicated and recreated through a transcranial implanted brain chip. For example, by modifying the Neuralink chip. That is why EM was invited to the gultik stall in Blok M today.

Prof HR’s next focus is on achieving self-sufficient vital energy in humans. This is partly inspired by the publication of Ryota Aoki’s team, which successfully integrated photosynthetically active algal chloroplasts into cultured mammalian cells towards photosynthesis in animals (Ryota Aoki et al. Proc Jpn Acad Ser B Phys Biol Sci. 2024).

Of course, this is also because vital energy molecules can now be mapped and understood for their biological and biochemical production processes.

Vital energy molecules are chemical compounds that function as energy storage, carriers, and sources used by living cells to carry out biological activities. This energy is needed for metabolic processes, molecular transport, biosynthesis, and muscle contraction.

The types include Adenosine Triphosphate (ATP), which structurally consists of three main components: Adenine, a nitrogenous base; Ribose, a pentose sugar; and 3 phosphate groups, linked by high-energy phosphoanhydride bonds.

The function of ATP is to store and transfer energy. ATP is hydrolyzed into ADP (adenosine diphosphate) and Pi (inorganic phosphate), releasing energy ~7.3 kcal/mol. This energy is used for protein synthesis, muscle contraction, and active transport.

Then there is Nicotinamide Adenine Dinucleotide (NADH and NADPH), which consists of two nucleotides linked by a phosphate group, adenine nucleotide, and nicotinamide nucleotide. NADH plays a role in oxidation-reduction reactions, transferring electrons in the electron transport chain. Meanwhile, NADPH is used in fatty acid biosynthesis and protection against oxidative stress.

Next, there is the FADH2 (Flavin Adenine Dinucleotide) molecule, which contains flavin (vitamin B2) bound to an adenine nucleotide. The function of this molecule is to carry high-energy electrons during the Krebs cycle and electron transport chain.

There is also GTP (Guanosine Triphosphate), which has a structural formula similar to ATP, but the nitrogenous base guanine replaces adenine. This molecule is used in protein synthesis and signal transduction.

Then there is Creatine Phosphate (Phosphocreatine), which consists of creatine and a high-energy phosphate group. This molecule provides quick energy to muscles through phosphate transfer to ADP.

The production process of this energy involves several mechanisms, including Glycolysis, which occurs in the cytoplasm, where glucose is broken down into two molecules of pyruvic acid, producing 2 ATP and 2 NADH.

Main Reaction:

C₆H₁₂O₆ + 2 NAD⁺ + 2 ADP + 2 Pi → 2 C₃H₄O₃ + 2 NADH + 2 ATP + 2 H₂O.

Then, of course, there is the Krebs Cycle (Citric Acid Cycle), which occurs in the mitochondrial matrix. Pyruvate is converted into acetyl-CoA, which enters the Krebs cycle, producing NADH, FADH2, and GTP.

Main Reaction:

Acetyl-CoA + 3 NAD⁺ + FAD + GDP + Pi → 2 CO₂ + 3 NADH + FADH₂ + GTP.

There is also the Electron Transport Chain and Oxidative Phosphorylation, which occur in the inner mitochondrial membrane. Electrons from NADH and FADH₂ are transferred through protein complexes. A proton gradient is formed and drives ATP synthase to produce ATP.

Main Reaction:

10 NADH + 2 FADH₂ + 6 O₂ + 34 ADP + 34 Pi → 34 ATP + 6 H₂O.

Specifically, in plants, there is the process of photosynthesis, which occurs in Chloroplasts, where light energy is used to produce ATP and NADPH. The light reactions produce energy molecules for the Calvin cycle.

Main Reaction:

6 CO₂ + 6 H₂O + light → C₆H₁₂O₆ + 6 O₂.

For utilizing energy reserves from fat, there is the Beta-Oxidation of Fatty Acids mechanism, which occurs in the mitochondrial matrix, where fatty acids are broken down into acetyl-CoA. Acetyl-CoA will enter the Krebs cycle to produce NADH and FADH₂.

Prof HR’s presentation up to this point has begun to provide insights into various possibilities that can be used as a basis for finding solutions to the crisis of humanity.

His thought framework is roughly as follows:

1. Taste Molecules and Their Physiological Effects can be replicated using BioPhysics methods, which can be stimulated on a chip like Neuralink.
2. Vital Energy Molecules can be reconstructed using Bioinformatics approaches in-silico and synthesized using de novo mechanisms in-vitro.

To synthesize molecules according to the current purpose, AI can help design with precision, optimal function, and validation. One way is by utilizing the capacity of Alphafold. AlphaFold is an artificial intelligence-based model developed by DeepMind to predict protein structures based on amino acid sequences. Accurate protein structures are key in designing molecules, especially for applications in pharmaceuticals, biotechnology, and molecular biology.

AlphaFold works by receiving the amino acid sequence of the target protein as input. This sequence provides basic information that determines how the protein will fold into a three-dimensional (3D) structure. This sequence is determined through techniques such as genomic or proteomic sequencing.

AlphaFold uses the Multiple Sequence Alignment (MSA) approach to compare the target protein sequence with a database of other proteins. MSA helps identify important evolutionarily conserved patterns for understanding the interactions between amino acid residues. Common databases used by AlphaFold include UniRef, MGnify, and PDB (protein data bank).

AlphaFold predicts the likelihood of contacts between pairs of amino acid residues in the protein using a transformer-based AI model, which helps determine the spatial relationships or distances between amino acid residues in the 3D structure. It also predicts torsion angles (phi and psi) that determine the orientation of side chains.

Therefore, this AlphaFold model is trained using a large dataset of known protein structures. It uses optimization algorithms to assemble amino acids into a 3D structure and employs techniques such as gradient descent to minimize free energy in the protein model, ensuring that the final structure is the most thermodynamically stable configuration.

AlphaFold validates the generated model by calculating the “confidence” value through metrics such as *pLDDT (predicted Local Distance Difference Test), which indicates the model’s confidence in the local structure prediction.

In the context of Prof HR’s idea, this importance is what led him to bring together the trio of YM, the molecular modeling expert from Bandung, with SA, the Transformer champion, and JH, the founder of NVIDIA, the producer of GPUs or brain networks for AI. Nvidia has various GPUs that can be used for AI development, such as:

Nvidia GeForce RTX, A GPU that can be used for various computational needs on different scales. Then there is the Nvidia DGX A100, this third-generation AI computing system is claimed to be the most advanced AI system in the world. DGX A100 has an AI performance of five petaflops and can be used to train AI systems with large datasets.

Then there is the advanced Nvidia H200, Nvidia’s latest AI GPU is equipped with HBM3e memory technology that allows the GPU’s memory capacity to be increased to 141 GB. There is also the Nvidia Blackwell B200, designed to process large data such as inferences from large language models (LLM).

Meanwhile, for aesthetic and artistic needs, there is Nvidia DLSS with AI-based rendering technology that allows it to reconstruct images to approach the visual quality of the original target.

The NVIDIA RTX 4090 is widely considered the best GPU for AI image creation. Its robust architecture, large memory capacity, and advanced tensor cores make it ideal for processing complex datasets and producing high-quality images.

To obtain the mechanism for synthesizing vital energy molecules and various proteins and fatty acids needed by the body, molecular modeling is performed, which is the expertise of Dr. Yusuf Muhammad from Unpad. Molecular modeling begins with the selection of the target molecule (protein, DNA, RNA, or small compound). This molecule is identified through laboratory experiments or computational predictions. The structure of the target molecule is input in a standard format (e.g., PDB for proteins). Data can be obtained through X-ray crystallography, NMR spectroscopy, or structure prediction (such as AlphaFold).

This process involves simulating the behavior of molecules under certain conditions, such as interactions with other molecules, structural changes, or responses to the environment. Simulating molecular models involves the laws of physics and quantum chemistry.

The model results are verified and adjusted with experimental data to increase accuracy, for example, using energy minimization algorithms. Some molecular modeling methods that are currently widely used include:

Molecular Mechanics (MM), where molecules are modeled as particles bound by interatomic forces. These forces are calculated using force fields, such as AMBER, CHARMM, or OPLS. Commonly used for simulating the dynamics of large molecules such as proteins and lipid membranes.

Then there is Quantum Chemistry (QM), which uses quantum mechanics to calculate the electronic energy and chemical interactions of molecules, for example, with Hartree-Fock (HF) to calculate the electronic wave function and Density Functional Theory (DFT) to calculate the electronic density distribution. Commonly used for studying chemical reactions, covalent bonding, and non-covalent interactions.

Then there is Molecular Dynamics (MD), which is commonly used to simulate the behavior of molecules over time by solving Newton’s equations of motion. Often utilized to model protein flexibility, conformational changes, and protein-ligand interactions.

There are actually a few more methods, but it seems that the methods and simulations above can already provide an idea of how and the principle of molecular modeling works. The relevance to Prof HR’s idea is that if we successfully model various vital energy molecules and also succeed in synthesizing them outside the body, whether through chemical or biochemical and biotechnological engineering, then the need for energy, protein, and various essential nutrients can be substituted with nano (10–9) capsules or tablets integrated with neurochips that can stimulate sensory areas with the taste spectrum we desire, even available in many options that can create a delicious sensation in our brains.

The implications?

1. The conversion of land for agricultural, plantation, livestock, and fishery needs can be reduced, even stopped.
2. The naturalization program for former agricultural, plantation, livestock, and fishery lands can bring large areas of green cover in the form of forests and ecosystems with very high biodiversity.
3. The greenhouse effect produced by the accumulation of greenhouse gases can be reduced, even completely eliminated, considering that apart from exhaust gases, the highest greenhouse gases are contributed by the livestock and agricultural sectors.

Thus, it is no wonder that Prof HR took the initiative to gather the four super scientists whose combined strength is hoped to produce a sustainable solution to the problem of the human population explosion and the carrying capacity of its environment.

But suddenly, I felt someone shaking my shoulder, and I woke up to find myself sleeping on a park bench under a Mahogany tree. Beside the bench stood Mbak Najwa Shihab, smiling. As I was still confused because I was suddenly ripped from my dream world, she asked, “Want some gultik?”

There, I truly lost my orientation and sense, staring intently at Mbak Najwa’s face with her prominent nose, I asked her, “Is this a dream within a dream?”

References

1. Anfinsen, C. B. (1973). Principles that govern the folding of protein chains. Science, 181(4096), 223–230. https://doi.org/10.1126/science.181.4096.223
2. Bartlett, G. J., Porter, C. T., & Thornton, J. M. (2002). The relationship of protein flexibility to function. Frontiers in Molecular Biosciences, 3, 1–20. https://doi.org/10.3389/fmolb.2002.00030
3. Brooks, C. L., Karplus, M., & Pettitt, B. M. (1988). Proteins: A Theoretical Perspective of Dynamics, Structure, and Thermodynamics. Advances in Chemical Physics, 71, 1–249. https://doi.org/10.1002/9780470141243.ch1
4. Case, D. A., Cheatham, T. E., & Kollman, P. A. (2005). Molecular dynamics simulations of nucleic acids using AMBER. Journal of Computational Chemistry, 26(16), 1668–1688. https://doi.org/10.1002/jcc.20290
5. DeepMind. (2021). AlphaFold: Revolutionizing Biology with AI. Retrieved from https://www.deepmind.com
6. Eisenhaber, F., & Bork, P. (1998). Prediction of potential GPI-modification sites in proprotein sequences. Protein Engineering, Design, and Selection, 11(12), 1155–1161. https://doi.org/10.1093/protein/11.12.1155
7. Ferrara, P., Apostolakis, J., & Caflisch, A. (2002). Evaluation of a fast implicit solvent model for molecular dynamics simulations. Proteins: Structure, Function, and Bioinformatics, 46(1), 24–33. https://doi.org/10.1002/prot.10016
8. Karplus, M., & Kuriyan, J. (2005). Molecular dynamics and protein function. Proceedings of the National Academy of Sciences of the United States of America (PNAS), 102(19), 6679–6685. https://doi.org/10.1073/pnas.0408930102
9. Levitt, M. (2009). Nature of the protein universe. Proceedings of the National Academy of Sciences (PNAS), 106(27), 11079–11084. https://doi.org/10.1073/pnas.0905029106
10. McCammon, J. A., Gelin, B. R., & Karplus, M. (1977). Dynamics of folded proteins. Nature, 267, 585–590. https://doi.org/10.1038/267585a0
11. Nelson, D. L., & Cox, M. M. (2021). Lehninger Principles of Biochemistry (8th ed.). W.H. Freeman and Company.
12. Orozco, M., & Luque, F. J. (2000). Theoretical methods for the description of the solvent effect in biomolecular systems. Chemical Reviews, 100(11), 4187–4226. https://doi.org/10.1021/cr000033j
13. Sali, A., & Blundell, T. L. (1993). Comparative protein modelling by satisfaction of spatial restraints. Journal of Molecular Biology, 234(3), 779–815. https://doi.org/10.1006/jmbi.1993.1626
14. Warshel, A., & Levitt, M. (1976). Theoretical studies of enzymic reactions: Dielectric, electrostatic and steric stabilization of carbonium ion in the reaction of lysozyme. Journal of Molecular Biology, 103(2), 227–249. https://doi.org/10.1016/0022-2836(76)90311-9
15. Zhang, Y., & Skolnick, J. (2005). TM-align: A protein structure alignment algorithm based on the TM-score. Nucleic Acids Research, 33(7), 2302–2309. https://doi.org/10.1093/nar/gki524