Team:Patras/Human Practices

Education

Integrated Human Practices

Everything started back in 2019! While attending the course of Molecular Genetics and Pharmacogenomics, we instantly realized that Pharmacogenomics, the study of how genes affect a person’s response to drugs, has great potential!

But, before fully forming our idea, we strove to understand the complexity of Pharmacogenomics and how famous it is in our community. Speaking with biologist experts, doctors, and pharmacists, we realized that pharmacogenomics would improve the clinical picture of a large subset of patients who suffer from drugs’ side effects. While we were excited to undertake this task and find the innovative solution, we first needed to find a suitable technique to implement. Using scientists and experts’ feedback, we comprehend that Artificial Intelligence can be used to achieve our goal.

We spoke with professional experts to pinpoint our detection approach, incorporating their thoughts to create an innovative method. Here, we highlight the most impactful interviews that we conducted and how our conversations shaped our project’s course.

Integrated Human Practices Timeline

February 2020

Being excited after receiving all this information from the course of Molecular Genetics and Pharmacogenomics, we decided to visit our professor’s office to find out what obstacles must be overcome for the application of Personalized Medicine into clinical practice. George P. Patrinos, professor of Pharmacogenomics and Pharmaceutical Biotechnology at the University of Patras and Head of the laboratory of Pharmacogenomics and Individualized Therapy, willingly guided us through our first steps. The scientists working in this specific laboratory are interested in ascertaining how Pharmacogenomics can reduce the overall healthcare expenditure by performing cost-benefit genomic analysis, facilitating the integration of Pharmacogenomics into clinical practice. During the discussion, he highlighted the three main problems that we have to overcome to achieve this integration. The high cost for genomic analysis, the exhausting time waiting for the results, and the non properly trained health professionals. He was the one who suggested us focusing on statins, a group of lipid-lowering medications, or antimalarial drugs, on the grounds that their metabolic procedure can be predicted more accurately.

George P. Patrinos

Through our conversation, we managed to define which were our project’s goals after underlining the barriers that Pharmacogenomics need to overcome to be implemented into clinical practice.

iGEM Patras Team

Mobirise

February 2020

We imagined making Pharmacogenomics accessible to everyone by eliminating these three barriers. After spending days studying the official guidelines for the administration of the aforementioned drugs, we attempted to discuss our dilemma with health professionals. Initially, to help ourselves choose, we got in contact with two significant cardiologists.

March 2020

Our country has numerous cardiology clinics staffed with renowned cardiologists that offer high-quality services in diagnosing, preventing, and treating cardiovascular diseases. Teams of experienced doctors work together daily to provide a reliable prognosis, diagnosis, and holistic treatment across the heart disease range, improving patients’ quality of life. Simultaneously, many cardiological clinics develop significant research activity in Cardiovascular Medicine, producing internationally recognized research work, and participating in global studies. After Professor Patrinos told us to analyze the metabolism of statins, a drug prescribed in cardiac diseases, we thought that a discussion with the head of the Cardiological Clinic of the University Hospital here in Patras MD George Chachalis would be extremely beneficial for us. Doctor G. Chachalis presented us that 40% of people die of cardiac diseases, so a method that can enhance the treatment of cardiac problems will be useful. Moreover, by hospital data, we see that doctors prescribe statins every day in plenty of patients, and 1 out of 3 end up with side effects, especially with muscle ache. Thus he believes that a genetic test would be beneficial and helpful, especially in the initial phases of cardiac diseases, to prevent more severe problems.

George Chachalis

By presenting us how many people suffer from such diseases, we realized that a project related to statins would greatly impact society.

Konstantinos Bakogiannis

We really took into consideration his advice to make our project as easy as possible to handle to have a greater chance to be established into the Public Health System.

March 2020

After the hospital of Patras’ cardiological clinic data about statins, we believed that we were on the right track regarding which one of the two above-mentioned drug categories we should select. To have a better perspective, we got in touch with another cardiologist, Konstantinos Bakogiannis, a member of the Hellenic Society of Cardiology. Having a Pharmacogenomics background, he deeply believes that its implementation into clinical practice would be revolutionary. He insisted that even though a project preventing statins’ side effects is necessary, there are a couple of problems that need to be faced in order to be integrated into the Public Health System. Thus, he urged us to “start with why” and make our diagnostic tool as easy to handle and efficient as possible. Finally, he advised us about the way we need to follow to promote our project to healthcare professionals.

March 2020

After completing our research, we returned to our professor’s office to discuss the findings with him. He agreed that working on statins, a clear-out drug category with only three genetic positions affecting their metabolism, is a brilliant idea. This drug category is widely prescribed to lower high cholesterol levels, which leads to an increased risk of coronary artery disease. He also proposed us using a portable laboratory to perform all the necessary experiments for genomic analysis. He insisted that in this way, Pharmacogenomics is becoming more accessible as it is cheaper for a physician to acquire a portable laboratory to perform such a test than to look for a physical laboratory or to create one in his place, something that it is proven difficult for someone settled in a remote area.

George P. Patrinos

Mobirise

Creating such a diagnostic tool available to not only metropolitan areas but also to the remote ones was one of his proposals. This is what made us look for a portable molecular laboratory to use in our project.

iGEM Patras Team

Mobirise

April 2020

While searching online for similar projects using portable laboratories, we found out BentoLab; a start-up, started as an iGEM Project of the UCL_E team back in 2013. This portable Molecular Biology laboratory contains all the needed instruments to accomplish a genomic analysis, but it has only been used before for educational purposes. We immediately realized that we could transform it into a diagnostic tool.

April 2020

It was about time for us to reach our next goal; making BentoLab as easy to handle as possible. One of the main difficulties that health professionals could face after completing the experiments would be interpreting the results into actual clinical advice. Thus, we thought it would be very beneficial for the user if there were software supporting him by translating the results into the proposed administered dosage. Our Dry Lab team suggested writing a code that uses either Computer Vision or Artificial Intelligence.

iGEM Patras Team

Mobirise

Ioannis Hatzilygeroudis

He inspired us to build our web-based application that uses AI software to understand the clinically useful information displayed on an agarose gel.

April 2020

To find out whether we should use Computer Vision or Artificial Intelligence, we decide to knock on an expert’s door in computer science. Professor I. Hatziligeroudis analyzed the benefits and drawbacks of each method. After further discussion, our Dry Lab team, responsible for all the technical issues, considered Artificial Intelligence a great idea to rely on to develop the automatic system. Thus, we discussed with Professor Hatzilygeroudis about the use of AI. He urged us to use deep learning methods, which could work better because the shapes we want to analyze are simple. Using the Random Forest program for classification and finding more data for analysis to have more valid results were also two pieces of advice that we deeply considered. Thus, we started running more experiments in the lab to generate all the photographs needed to train our Machine-Learning Algorithm.

April 2020

Another step towards completing our BentoLab-AI system was to evaluate if health professionals would use it. Diamantis Klimentidis, a clinical pharmacist in the greek public hospital “Agia Aikaterini,” offered to help us with that. After presenting to him our concept, he made us understand that genomic tests are not as simple as we think to be performed by any health professional due to lack of training. “Try to make it as easy as possible for everyone to use it,” he quoted, “is what will skyrocket your project.” He advised us to continue our project by making it a key turning solution to be more specific with Greek hospitals and doctors and calculate the cost-benefit effect to convince that our innovative idea is worth it.

Diamantis Klimentidis

Trying to follow his advice, we developed our step-by-step guidanceto assist even inexperienced health professionals in performing genetic analysis.

3 Medical Students

The “tests” were performed on three different days in order to have the opportunity to build on each feedback and make the instructions as understandable as possible by any user.

May 2020

To ensure that our BentoLab-AI system provides users with reliable assistance, we followed the trial and error method. That’s why we invited to our lab three medical students to have it tested. Their feedback was constructive, as they let us know whether the provided information was clear enough or there were modifications to be made.

July 2020

Having created a functional system, we would like to test it’s after iGEM potential and what we were missing in entrepreneurship. And there was no other better advisor than BentoLab; the start-up whose portable laboratory we are based on. We believe that it will be beneficial to discuss with Philip Boeing, one of the constructors and co-founders of BentoLab company, to advise for iGEM competition and Bento Lab’s use. He urged us to think about our project’s communication early on and promote our work on social media and websites regularly. Moreover, Philip encouraged us to prioritize the ideas if we want to make an excellent result. That’s why, for the first time, we approached especially cardiologists and not generally doctors. So we keep working with Bento Lab and thinking of how we can together find a manner to enhance the machine without upsizing it a lot.

Bento Lab

Mobirise

The two pieces of advice we ultimately kept from our conversation with Philip was to spread our idea before it is too late and pay attention to ethical and legal issues.

Vasileios Karkatzounis

His assistance was vital for creating this consent form legally accepted in Greece.

August 2020

Finding a specialist lawyer to guide us through our future steps wasn’t as easy as it sounds! Vasileios Karkatzounis, a legal engineer specialized in the proper use of Artificial Intelligence, highlighted all the legal obstacles we may face while trying to turn our project into a sustainable business plan. First of all, he informed us about the regulatory system and guidelines that were published by the European Council in 2018. Thus, he urged us to read them through, especially those referred to personal data used for the AI system. To be more specific, he told us that it is necessary to make a consent form to be legally valid because we use health data for developing an AI system. After his advice, we created a form that consists of the reason, the aim of using these data, the duration, the study results, the owner, and the legal base system of our system. Last but not least, encourage us to use these data to train the algorithm, so if we want to prevent mistakes, we have to be sure about the accuracy and validity of them.

September 2020

We could not skip over the delicate and severe issue of Bioethics. Athanasios Panagiotou, a member of the European association of health law, undertook the challenge of helping us with this. His contribution was crucial since he mentored us on finding the best way to promote our project to prove that it does not pose any risk for our society. For this reason, he proposed us to separate the instruction that will be extracted by the algorithm, in high and low-risk levels, to diminish the unexpected errors. He also highlighted the importance of using the BentoLab- AI system as an assistant diagnostic tool and not as a doctor’s replacement. To prevent any problems with other doctors such as microbiologists, he told us to analyse the need for such a technique in remote areas, where only one doctor is in the health center.

Athanasios Panagiotou

The most purposive advice was to clarify that the AI system will be used for doctors’ advisory services. It won’t have the crucial responsibility of the dosage adjustment and won’t be used as a doctor’s replacement.

About

16 students from Patras blending Pharmacogenomics with Artificial Intelligence to redefine medicine

Follow Us