How This Neovarsity Alumnus Is Advancing Drug Discovery with Machine Learning

Born and raised in the vibrant cultural crossroads of Morocco, Yassir Boulaamane has journeyed from being a biology undergraduate in Rabat to a soon-to-be PhD in Computational Chemistry in Tangier.

His academic path, deeply rooted in Morocco's educational institutions, took an innovative turn when he discovered Neovarsity.

Here, he acquired cutting-edge cheminformatics and machine learning skills, which eventually became central to his research.

Now, as a young Moroccan researcher, armed with this unique blend of expertise—a solid foundation in computer-aided drug design supported by these advanced skills—Yassir actively applies his knowledge to innovate on early-stage small-molecule drug discovery.

He has recently published research on antibiotic discovery for Acinetobacter baumannii infections, beautifully demonstrating how he applied his newly acquired machine-learning skills to real-world drug discovery challenges.

In this study, Yassir and his co-authors used a multi-step machine learning-based virtual screening approach, successfully identifying and experimentally validating demethoxycurcumin as active against all tested A. baumannii strains, both in monotherapy and in combination with colistin.

Read the full research here.

Notably, A. baumannii poses a severe global health threat due to high levels of antimicrobial resistance, leading to substantial morbidity and mortality (e.g., 26% to 68% in the USA).

Novel strategies beyond traditional drug discovery of antibiotics are imperative.

Yassir's findings highlight the potential of machine learning in discovering effective antimicrobial agents against A. baumannii infections, offering a promising strategy to address antibiotic resistance.

I sat down with Yassir to learn more about his journey, his new research, his career inspirations, and how he plans to implement these skills further.

Hi Yassir, thanks for chatting with me today! To start, can you share a bit about your early academic ambitions and any challenges you encountered along the way?

In middle school, I dreamed of becoming a pharmacist, which led me to pursue biology at Mohammed V University in Rabat. My journey hit a roadblock when I failed a crucial exam in my first two years. This setback was devastating, but it also became a turning point. Instead of giving up, I channeled my disappointment into determination. I immersed myself deeply in biology, becoming intensely dedicated to mastering the subject. I focused on thoroughly understanding the core principles and expanding my knowledge across various biological disciplines. This experience taught me resilience and turned a major challenge into an opportunity for growth and a deeper passion for my studies.

That sounds like a significant challenge. How did your interest in computer science develop while you were studying biology?

My interest in computer science developed alongside my biology studies. While pursuing my biology degree, I taught myself programming languages like HTML and basic Python out of personal interest. The real intersection came during my master's in Biology and Biomaterials, where I encountered molecular modeling. This field fascinated me as it combined biology with computational tools, allowing for visualization of complex biological processes. As a visual learner, I found this approach incredibly effective. This experience was pivotal, showing me the powerful combination of biology and computer science, and ultimately shaping my career path towards the intersection of these fields.

That’s fascinating! How did studying molecular modeling during your master’s lead you to start working on computer-aided drug design?

My transition to computer-aided drug design was a natural progression from molecular modeling. During my master's thesis on MAO-B inhibitors for Parkinson's disease, I applied computational tools to a real drug discovery problem. The pandemic lockdown accelerated this shift, allowing me to focus intensively on dry lab work from home. This experience demonstrated the power of computational methods in drug discovery, cementing my interest in the field. It led me to pursue a PhD, where I've continued to develop my expertise in computer-aided drug design, bridging biology, chemistry, and computer science.

What got you interested in using machine learning for drug discovery?

My interest in machine learning for drug discovery arose from the need to accelerate the identification of therapeutic natural products for neurodegenerative diseases. Traditional methods struggled with the vast databases of compounds available. Machine learning offered a solution to rapidly analyze this complex data, identifying potential drug candidates much faster. Its ability to handle multidimensional data and uncover novel patterns made it particularly appealing for drug-target interactions. This combination of efficiency and potential for new insights made machine learning an exciting and logical evolution in my research approach.

Related readings: What is data-driven drug discovery and how is it transforming traditional approaches?

How did you first discover Neovarsity, and what attracted you to our programs?

I discovered Neovarsity through a Facebook post in the Bioinformatics, Cheminformatics & Computational Biology community. What caught my attention was a workshop on machine learning in drug discovery. The workshop's focus on practical applications really appealed to me, especially the opportunity to work with open-source bioactivity data like the ChEMBL database. I was particularly interested in learning how to build QSAR models for predicting drug activities, as this aligned perfectly with my background in computational chemistry. Neovarsity offered the hands-on experience I was seeking to advance my skills in this field.

It’s great that you found us through our Facebook community! Which specific courses or modules at Neovarsity had the biggest impact on your research in ML-driven drug discovery?

The Advanced Machine Learning for Drug Discovery course had the biggest impact on my research. It provided hands-on experience with QSAR modeling, crucial for my work. The course stood out by emphasizing critical thinking and addressing real-world challenges. I learned to tackle data imbalance and bias, interpret models mechanistically, and identify chemical features contributing to drug activity. This comprehensive approach enhanced my understanding of both the potential and limitations of ML in drug discovery, significantly shaping my current research direction and improving my ability to apply ML effectively in this field.

How did you integrate machine learning with your existing skills in molecular modeling?

I integrated machine learning with molecular modeling by combining QSAR models with docking and dynamics simulations. ML efficiently screened and prioritized compounds, while molecular modeling provided detailed insights into target interactions. For instance, we'd use ML to predict compound activity, and then apply molecular docking to study how promising candidates bind to targets. This approach leveraged ML's speed and pattern recognition alongside molecular modeling's atomistic-level analysis, significantly enhancing our efficiency in identifying and analyzing potential drug candidates.

That’s a smart approach! What made you decide to use machine learning for antibiotic discovery?

I chose to use machine learning for antibiotic discovery to address the challenges posed by A. baumannii, such as its rapid development of multi-drug resistance and highly impermeable outer membrane. Machine learning offered a way to accelerate the discovery process and improve our chances of finding novel compounds. By building QSAR models from known A. baumannii inhibitors in the ChEMBL database, we could leverage existing data to predict the activity of new molecules.

What were the key advantages of using AI in this study compared to conventional drug discovery methods?

AI helped us prioritize the most promising compounds more efficiently than traditional methods. By narrowing our focus to a subset of molecules with a higher likelihood of success, we saved time and resources and increased our chances of discovering antibiotics with new mechanisms of action. Although machine learning was just one part of our strategy, it provided a strong starting point and guided our research more effectively.

That’s impressive! Were there any specific skills or techniques from Neovarsity that helped with your published work on antibiotic discovery?

The curriculum at Neovarsity was incredibly valuable for my work in antibiotic discovery. The data preprocessing techniques I learned there were essential because chemical data can be quite complex. Neovarsity taught me how to clean, normalize, and prepare molecular data effectively. I also learned to build and fine-tune machine learning and deep learning models using algorithms like random forests, support vector machines, and convolutional neural networks. These skills were key to advancing my research and making my published work possible.

What were the most significant outcomes of your ML-driven approach in this study?

The most significant outcome of our ML-driven approach was discovering the curcuminoid class among our top hits. This class exhibited the highest affinity for outer membrane protein W (OmpW) in A. baumannii, which is a promising target for new antibiotics. Specifically, we identified demethoxycurcumin as a standout compound. Our models and subsequent tests showed that demethoxycurcumin binds with high affinity to OmpW and demonstrates impressive antibacterial activity against A. baumannii, surpassing previous reports for natural products in the literature.

That’s a notable finding! How do you think your ML-based approach could influence the wider field of antibiotic discovery?

This discovery not only validated our ML-driven approach but also provided a promising lead for new antibiotics. It demonstrated how machine learning can identify novel compounds and targets that might be overlooked by traditional methods, potentially transforming the field of antibiotic discovery.

That’s exciting to hear! What are your future plans in the field of drug discovery?

I am seeking a postdoctoral position to advance drug discovery through the application of AI and machine learning, with a particular interest in antibiotic development. My research focuses on developing robust, interpretable models that bridge academic and industry approaches, leveraging my expertise in molecular modeling and data analysis.

I aim to enhance model quality, applicability, and transparency through techniques such as mechanistic interpretation and structure-activity relationship analyses. I am eager to collaborate with multidisciplinary teams to drive innovation and ultimately contribute to the development of novel therapeutics that address unmet medical needs.

Thanks for sharing your story with us, Yassir. That sounds promising! I wish you lots of luck in your growing career. Before we wrap up, one last question: Would you recommend Neovarsity to other researchers? If so, what makes it a great choice?

Certainly, I highly recommend Neovarsity to other researchers in the field of drug discovery and computational chemistry. My experience there was invaluable for several reasons.

First, the curriculum is extremely well-tailored to the needs of our field, offering a perfect blend of theoretical foundations and practical, hands-on experience with the latest tools and techniques in machine learning and AI as applied to drug discovery.

Second, the instructors are not just academics but also industry professionals with real-world experience. This means you learn not only the theory but also how to apply these concepts to solve actual problems in drug discovery.

The networking opportunities are fantastic. You are surrounded by like-minded researchers from various backgrounds, which leads to interesting collaborations and exchanges of ideas.

Lastly, the skills I gained at Neovarsity directly translated into my work in antibiotic discovery. The techniques I learned for data preprocessing, model building, and result interpretation were crucial to my research success.

For anyone looking to leverage AI and machine learning in drug discovery, Neovarsity provides an excellent foundation and can significantly accelerate research capabilities.

Note for Recruiters

For recruiters and PIs interested in a dynamic professional like Yassir, feel free to reach out. Yassir is a hardworking individual with a quick learning ability and an explorative mindset.

He is currently open to postdoctoral positions or roles in the industry. For inquiries, please contact us at [email protected] or connect with him directly via LinkedIn. Yassir brings a valuable blend of expertise in computational chemistry, programming, and extensive hands-on experience with small molecules through Neovarsity's Molecular Machine Learning programs.

—

You can also explore his profiles to see more of his work:

Research Profile: https://yboulaamane.github.io/

GitHub: https://github.com/yboulaamane

LinkedIn: https://www.linkedin.com/in/yassir-boulaamane/

Note for Learners

If you're wondering if learning machine learning for drug discovery is a good move for your research and career, Neovarsity is here to help.

Attend one of the upcoming free workshops to get a taste of what it entails, or check out these blog articles to learn more about this emerging, high-level skills field:

11 reasons to master cheminformatics before machine learning in drug discovery

9 applications of machine learning in drug discovery

10 ways AI is transforming drug discovery in 2023: A guide

You can also explore Neovarsity's TechBio curriculum and book a call with an expert advisor to discuss your options and find out if it is a good fit for you.

Molecular Machine Learning Foundation

Advanced Machine Learning for Drug Discovery

AI for Drug Discovery: Deep and Reinforcement Learning

We offer a 6-9 month specialized curriculum designed to help you gain advanced skills in applying machine learning and artificial intelligence to small-molecule drug discovery and build a successful career in this field. Check out the Molecular AI Specialization.

If you're curious to learn more about learner stories, check these out:

Meet Stefani: A Paris-Based Quantum Chemist Exploring Molecular AI with Neovarsity

Meet Rutgers Medicinal Chemist Who’s Forging a Path with Molecular AI at Neovarsity

Meet the Toxicologist Who’s Forging a Career Path with Molecular AI

Meet Sahil: A Norway-based Computational Chemist Developing Molecular AI Skills at Neovarsity