SOCIAL SUSTAINABILITY

AI donates to research into a rare disease

We started in 2014 as an innovative start-up, becoming in 2019 an innovative microenterprise, with the objective of helping companies to evolve using Artificial Intelligence and Machine Learning. Although we are a small team, our aim is that our ideas and actions can contribute to social and environmental well-being.
We believe that our expertise, acquired over our 6 years of activity, can be of help if applied to the study of pathologies, with the aim of supporting research and helping patients to receive customized diagnoses and treatments.

 

AKCAPTONURIA
In March 2020, in the height of the covid-19 pandemic, we decided to dedicate our time, resources and energy to research into a genetic disease, alkaptonuria (AKU). Few people know about it; we had never heard of it before the project. We discovered that it’s an ultra-rare disease (affecting around one person every 250 thousand – 1 million) and, as of today, it’s not clear why a person is born with this disease. The body of a person affected by AKU is unable to dispose of homogentisic acid (HGA), which accumulates in the joints, heart and connective tissues. The consequences are blackening of the nose, ears and joints, pain and difficulty in moving and heart problems. There is no cure, but patients receive supportive care with painkillers and anti-inflammatories. What the doctors do know, however, is that the course of the illness is the same for everyone.

 

THE CHALLENGE
Professor Santucci,
Head of the Biotechnology, Chemistry and Pharmacy Department in the University of Siena, proposed to us a very challenging project: to apply Artificial Intelligence to small data in order to obtain a more objective method for quantifying the course of the disease. It wasn’t the first time that we had to deal with small data.

 

BIG OR SMALL DATA?
We’ve got used to giants like Google or Amazon applying AI to enormous quantities of data. Big Data is just one area that AI can be applied to, but is not necessarily indispensable for innovation. We are convinced that every company, be it small or large, can use its data to embark on a path of innovation and integrate modern analytical techniques. Hence our slogan: “Big or small data: many, few, whatever is needed!”. What makes the difference is knowing how to apply the best methodology and technique for those data: the alkaptonuria project was an opportunity to demonstrate this.
 
By crossing the data that we have regarding 203 European patients (regarding age, sex, country of origin, behavioural information and any other pathologies, such as diabetes, hypertension, etc.) with questionnaires compiled by the patients, in which the symptoms of the disease are quantified through the Quality Of Life Score (a score from 0 to 100), we managed to obtain two results to deliver to the University of Siena.

 

Using Artificial Intelligence for predicting the state of progress of the disease on new patients:
In the first phase we wanted to predict, on new patients, quality of life and understand the state of progress of the disease.

 

The identification of correlations between drugs and disease:
In the second phase we wanted to understand if there was a correlation between the use of drugs and the disease. Every drug is obviously composed of different chemical elements. We grouped the drugs into 33 categories (anti-inflammatories, painkillers, opiates, etc.) on the basis of main active ingredients. Although there weren’t a lot of data, thanks to advanced statistical analyses we identified 8 types of drugs that have an impact on the disease.

These results are not the end of the process, but a new beginning of research in the medical field; they are the demonstration that also small data can make the difference.

Personal story
— the interview video–

  1. The project carried out together with the University of Siena was somewhat of a gamble. Besides the difficult theme, you had to work in a period that was also problematic. This experience, however, has reinforced your conviction that Big Data are not necessarily essential for obtaining results. Do you want to tell us about your experience?

At the beginning of March, when Italy was in lockdown for the pandemic, we decided to donate our Data Science expertise to support research into this rare genetic disease. Working from home wasn’t a problem; almost all our projects are carried out remotely. We spent 4 weeks studying alkaptonuria, a disease which we knew nothing about; thanks to Annalisa, her team and research of the relevant literature, we learned about the world of rare genetic diseases and the support that Data Science is able to give in carrying out an analysis when there are little data.
We were used to thinking that it was possible to apply AI only to enormous quantities of data but using Big Data is only one of the possible approaches, luckily. We had a dataset with information on only 208 people. Once we’d cleaned the dataset of incomplete information, only 138 people remained.   
We knew that the data available were few, but we are convinced that every company, be it small or large, can use its data to embark on a path of innovation and integrate modern analysis techniques. Hence our slogan: “Big or small data: many, few, whatever is needed!”. What makes the difference is knowing how to apply the best methodology and technique for those data: the alkaptonuria project was an opportunity to demonstrate this. After 3 months of work, we had obtained results useful for the research, which will be taken forward by the Biotechnology, Chemistry and Pharmacy Department of the University of Siena

  1. Besides great expertise in interpreting numbers and using algorithms, what is your point of strength in facing such complex challenges?

As always, it’s reciprocal cooperation and the diversity of expertise that allow for the success of Data Science projects.  We are so convinced of this that our team has been specifically selected to reduce bias and to be able to interpret complex situations from different points of view. We have formed a team which varies in terms of age, sex, culture, ethnic origin, education and background; the diversity that characterises us has helped us to face challenges that others have refused in the past. The study carried out regarding alkaptonuria was no less complex than other projects but, confident in our expertise and experience, we accepted the challenge and obtained good results that can in the future can help in the choice of support treatment and in the development of targeted drugs. .
Convinced of this, for the alkaptonuria research project we tried different AI techniques and advanced statistics, comparing the results obtained with Annalisa’s team. Our expertise was combined with the skills of Annalisa’s team in the validation of the result, and together we achieved results useful for continuing the study and research of this rare disease.

  1. Technology today allows us to unite the most brilliant minds, wherever they may be. Is that how it’s been for you?

Yes, having a widespread structure, we are used to working remotely.  Except for certain cases where, due to the sensitivity of the data, it’s preferable to work on the customer’s premises, our projects are all carried out remotely. This work method allows us to reach anyone, without limits. Whoever begins a collaboration with us through an experimental research and development project, like this one, provides a subset of data in order to evaluate the feasibility of the project.  We experiment and propose different analytical techniques and together with the customer we choose the best path to take. Several years of experience in Data Science projects have helped us to refine an effective working method. The most important challenge of the project was the small data, but this didn’t prevent us from completing the work and we tackled every phase with scientific method, conscious of having achieved a useful result for the study of the disease.  Our experience and our work method have given rise to Ilyde, the platform where it’s possible to develop Data Science projects. It’s the first Made-in-Italy work environment in which data scientist can collaborate remotely, using a shared place where to work, keeping track of the tests and parameters of the models. Ilyde has enabled us to continue our work and not to be too affected by the lockdown; we have continued to manage our projects remotely, both with colleagues and customers, and so we were able to start immediately on the alkaptonuria project.

 

  1. The small amount of data available makes research on this disease difficult, but technology has come to our aid. When did the Biotechnology, Chemistry and Pharmacy Department of ethe University of Siena cross Hopenly’s path and what was the connecting link that made this research possible? –Annalisa–
  2. There’s still a lot of misunderstanding about the use of AI, yet it could offer many benefits in the field of healthcare. What extra benefit has Artificial Intelligence given you with respect to other analytical methods and what will the next steps be? –Annalisa–
  3. Rare diseases, unfortunately, seem to be relegated to second place, but you have shown that it doesn’t have to be that way. Research never stops, and every day technology and science proceed hand in hand towards scientific progress. Why is it important to work on a rare disease? — Annalisa —