Protein Designing and Artificial Intelligence: Demystifying 2024 Chemistry Nobel Prize

Content Number: 22
Author Name: Rabbia Shoukat
Author I’d: SBPWNC – A22
Educational Institution: School of Chemistry, University of the Punjab, Lahore, Pakistan
Content Title: Protein Designing and Artificial Intelligence: Demystifying 2024 Chemistry Nobel Prize

“In stone age, people learn how to control metal and other artefacts

In industrial revolution, people learn how to handle things via steam engines

 In digital revolution, people comprehend how to stockpile information digitally

In protein designing revolution, we are on the way to unravel how to control biomolecules in ways not possible before”

                                                               ~  David Baker – 2024 noble laureate in Chemistry

Introduction

Protein designing, along with the prediction of protein structure, sparks the world with key ideas to get rid of the menace of many biological abnormalities, such as cancer and viral diseases. However, the retrieval of artificial intelligence (AI), in the process of prediction and designing, compels the trio – Baker, Hassabis, and Jumper – to put the victory of 2024 Noble Prize in their bucket. The current stride is not less than a quantum leap in the world of biochemistry. First, the designing of protein structure by Baker is one of the magical steps. He involves different institutions and people at home to volunteer through Rosetta@home – computer volunteer program – which actually aids in his work of protein designing as Baker can’t afford too much computers and space for carrying out the job. Second is the prediction about structure of protein through AI algorithm. Hence, AlphaFold2 proves to be a breakthrough to decipher the mystery of protein structures. Both events take science many leaps forward to aid humanity.

The 3D model has been released from Baker’s lab, showing nanomaterial upon which upto 120 proteins can be cringed together, is a testament to his unrelenting endeavour in complementing the designing process via applying AI algorithms.

Animations: ©Terezia Kovalova/The Royal Swedish Academy of Sciences

Contextualising the worth of Protein models

“Protein – the elementary units of life; hence, life cease to exist without protein”.

Only one factor can’t gauge the significance of protein in carrying out various chemical processes supporting life in living creatures. However, there are multitude of aspects, including their structure, shape, unique arrangement, and distinctive sequence of amino acids, that decide what will be their functioning. Four basic structures are reported for protein, and out of those, secondary configuration carry out the prominence in Baker’s research. Moreover, Baker work on alpha helical secondary structure. In one of his interviews, Baker intimates about the much complexity of beta sheet as one of the viable reason of not opting it for computational work.

Protein Designing and Artificial Intelligence

The process of protein structuring with the application of AI proves to be a breakthrough in the world of biochemistry. AI revolutionises the process of model designing, and Baker, undoubtedly, was the one of the personalities, who rightly reap its benefits. Necessary to mention is their AI program with the name of Rosetta software. Rosetta converges short fragments of protein from protein unrelated structures having sequence close to that of local data bank and then it is optimized for that of target. Hence, AI plays unmatchable role in protein designing; hence reducing the time, amplifying the output, and posing unexplainable gain to the humanity at large.

It is the Rosetta software program, which is an AI-driven algorithm, letting wonders in designing process. However, it is not just factor, behind the curtain is the brainchild of the David Baker, who contemplate to do something new, and in this quest, he resorted to de novo designing of enzymes [i]with enhanced power of catalytic processes. Baker formulates such a protein that clutches to steroids with proven selectivity, as well as sensitivity. The basic process of designing protein for cancerous cells is aptly illustrated below with diagrammatic stepwise representation.

Designing Proteins for Cancerous Cells

Protein designing technique, supported by artificial intelligence (AI), provides a comprehensive and quicker way to design hundreds or even thousands of models at the concurrent time. According to the reported data by Baker, the first draft takes almost 4 to 6 weeks and then depending upon the results, designs are updated. However, the process is automatically carried out in various steps. First, identification of binding site on the targeted cell occurs followed by designing of candidate protein. Candidate protein must be rough in shape, as well as complementary to the target. After it, sequence on the binding interface is checked, which is chemically and geometrically engaged with the target. Further process is vividly shown in the below protein designing steps.

Apart from the designing, predictability of structure of protein is also the step worth-explaining carrying equal weight in winning Nobel prize. The structure prediction falls under the ambit of Hassabis and Jumper’s work.

Prediction of Protein Structure – Hassabis and Jumper’s Contribution

Hassabis and Jumper put immense contribution in forecasting protein structure. This is an effort worth- celebrating, which led them to win half of the Nobel Prize. Prediction about protein structure means having no prior idea about it. According to the researches, there are almost 500 distinct proteins having unique sequence and well-known 3D structure that can be employed to gauge the extent upto which technology estimates protein structure[i]. Protein structure, undoubtedly, is of utmost importance, and incapability to uncover it poses much implications. Furthermore, protein new arrangements arise as an output of folding of previous structures. For instance, twisting of primary structure led to secondary, folding of secondary structure led to tertiary, and intermingling of two or more polypeptide strings led to quaternary structure. To add in it, it is crucial to discuss the significance of protein folding. One the one side, folding decides protein functionality, and on the other side, misfiring led to multitude of diseases, such as Alzheimer and Parkinson. However, breakthrough in the AI world simplify the complexity of the task being done, and the conundrum is sort-out through breakthrough achievement – the AlphaFold2 breakthrough.

Unravelling the Puzzle of Protein Folding – AlphaFold2 Breakthrough

1. Data Input and Database Searches

AlphaFold2 has familiar protein structure, as well as amino acids (AAs) sequence in the database. When the unfamiliar protein structure and the unknown AAs sequence is fed into the AlphaFold2 then processing starts.

2. Sequence Survey

Amino Acids chains which are conserved during the process of an evolution are aligned by AI model and then the algorithm explores about the co-evolving AAs. Interaction occurs in 3D structure and the interaction is in between charges of opposite nature. It is elaborated in the diagram given below.

3. AI Survey

AI model employs transformer, which is a neural network, detects the important structures. AlphaFold2 refines the chain, and the process go on to the next sequential step.

4. Hypothetical Structure

AlphaFold2 merges the mystery of provided unknown AAs, and different routes are tested for formulation of an imaginary structure. It is clearly represented in the below structure that after three successful cycles, AI software lands at a particular hypothetical structure.  The credit of diagram, given below, goes to the publishing website mentioned on the picture.

One aspect needs to be mentioned is that the prediction of protein structure lags behind protein designing because for designing, target is idealised and its probability amplified much after optimization of targeted. In contrast, there is no prior information in case of protein structure prediction.

Applications and Future Prospects

The recent ground-breaking innovation in the realm of protein designing aided by artificial intelligence, as well as anticipating protein structure, opens up doors to the extensive applications in various scientific fields. Above all, detection of chemicals – like fentanyl – in the environment, formulation of molecular rotor, production of vaccines, like Influenza vaccine, and the creation of protein tiny sensors are worth-mentioning. The actual demonstration from Baker’s – noble laureate – laboratory is mentioned in the figure given below. Apart from the fact that recent computational designing of protein has innumerable applications, but there are few whose mentioning is of utmost significance. First, Fentanyl, which is an artificial opioid, is more pernicious than that of morphine and heroine. Fentanyl is processed in foreign labs secretly. According to the investigation, the opioid is smuggled to the United States of America from Mexico and is vend off in drug market illicitly.[i] Second, formulation of protein molecular rotor defends the credibility of Baker’s work. Molecular rotor computes protein assemblage. Its significance is elaborated from one of the researches that show how protein aggregation provides useful insights related to structural alterations. [ii]

Animations: ©Terezia Kovalova/The Royal Swedish Academy of Sciences

Third, protein designing provides a road-map for the vaccines production. For examples, Influenza virus vaccine is deciphered through designing protein, imitating for Influenza virus. As shown in the picture given below, nanoparticles (yellow coloured) encapsulated by protein structures that simulate for Influenza virus, and it serves as a potential vaccine for Influenza. Its successful application is also performed in animal model. Fourth, tiny protein sensors have been created, which serve as a geometrical shaped protein in which external influences bring about shape changes. Owning to the specialty, tiny sensors find ample applications in the fields of neuroscience and research. Furthermore, scientists are now on the way to design such enzymes, which can tackle the menace of plastic pollution by plastic degradation.

©Terezia Kovalova/The Royal Swedish Academy of Sciences

Conclusion

In conclusion, the work of Nobel laureates marvel the research area. Artificial Intelligence, which aids humans in carrying out tasks from simpler to multiplex, finds no exception in protein designing and prediction. It is AI, which has turned the Hassabis and Jumper’s 50 years old dream into reality. Owning to the landmark victory, scientists can now deal with life-threatening abnormalities. Above all, cancer tops the list in this regard. Moreover, vaccine for Influenza virus is also a step worth-quoting that become easier just because of AI applications. However, it’s not the end; still, much is awaited to be explored in the ocean of mysteries. What is needed – a curious mind with the courage to take up the task, and then let it done. As David Baker during his work of designing protein aptly penned down as:

“We show them the successful magic-like designs, but seldom talk about all the steps and the things that didn’t work”.

                                                                                                 ~David Baker

References

i https://www.nobelprize.org/prizes/chemistry/2024/press-release/

ii https://pmc.ncbi.nlm.nih.gov/articles/PMC8289003/

iii https://pmc.ncbi.nlm.nih.gov/articles/PMC8289003/

iv https://www.linkedin.com/newsletters/texploration-with-david-cain-7075138327997231105

v https://www.embl.org/news/science-technology/alphafold-wins-nobel-prize-chemistry-2024/

vi https://www.dea.gov/resources/facts-about-fentanyl#:~:text=Illicit%20fentanyl%2C%20primarily%20manufactured%20in,on%20the%20illegal%20drug%20market

vii https://www.sciencedirect.com/science/article/pii/S0167732224024097

viii https://www.nobelprize.org/uploads/2024/10/fig4_ke_en_24_2.pdf

Previous Blog: Phytochemical Analysis and Antimicrobial Activity of Cinnamomum Verum

Follow Us on

FACEBOOK

INSTAGRAM

YOUTUBE