By Emma Anderson, Apostolos Gkatzionis, Lucy Goudswaard, Ruth Mitchell, Chin Yang Shapland, Kaitlin Wade and Venexia Walker
In recent years, there has been an explosion of research in Mendelian randomization (MR). It is a useful tool for inferring causality between exposures and outcomes of interest using genetic data. The Mendelian randomization conferences organised by the MRC Integrative Epidemiology Unit (IEU) regularly provide an opportunity to explore these issues.
In 2018, one of the conference attendees was Dr. Sarah Atkinson from Kemri Wellcome Trust in Kilifi, Kenya. As the conference progressed, she began chatting with Dr Ruth Mitchell and the idea emerged of IEU hosting an MR course for African scientists. Sarah applied and was awarded funding from the Wellcome Trust, which was supplemented by the IEU, to host six experts in MR from the IEU to teach a five-day course on Mendelian randomization to African researchers in Kilifi.
The course, which took place in November 2022, included 22 participants, who were selected from over 300 applicants, from eight countries across Africa (the Gambia, Nigeria, Sudan, Ethiopia, Uganda, Burundi, Kenya, and Tanzania). The course participants had a wide range of research interests, including malaria, mycobacterium tuberculosis, sleeping sickness and neurobehavioral outcomes.
The intention was for these participants to not only attend and learn from the course but also to network to stimulate collaborations between research groups in different African countries and between the UK and Africa.
Day 1 of the MR workshop in Kilifi Kenya opened by the KEMRI-WELLCOME director. Interesting first lecture on GWAS. #GEfAS. pic.twitter.com/VlpPOFpz5S
— Jarra (@JarraJmanneh) November 14, 2022
Developing skills to implement MR
The aim of the course was to teach the participants how to implement MR, and how to use the IEU-developed and open-source MR-Base software. Our hope was that, by the end of the course, they could apply MR to their own research. Specifically, in addition to knowing how to implement MR, the intended learning outcomes were to:
- Describe the basic concepts of genetics, particularly related to genome-wide association studies;
- Explain the principles and assumptions of MR;
- Discuss the strengths and limitations of MR for addressing causal questions;
- Critically appraise MR papers, undertake MR analyses and interpret results; and
- Describe the challenges of applying MR to molecular traits and drug targets.
As the course was based upon both the Genetic Epidemiology and the MR short courses at the University of Bristol, most of the teaching materials were readily available. However, a couple of sessions were added to tailor to the tutors’ expertise and to the experiences of the participants. Dr. Venexia Walker, Dr. Kaitlin Wade and Dr. Lucy Goudswaard specialise in using MR for predicting drug target effects, the gut microbiome and molecular traits, respectively. Dr. Kaitlin Wade also created a bespoke practical on undertaking genome-wide association studies within a setting that did not have access to high performance computing to participants with diverse backgrounds and experiences.
It’s the start to an exciting week of Genomic epidemiology for African Scientists #GEfAS #training at the Kilifi coast. The 1st Mendelian randomization #MenRan conference in Africa. pic.twitter.com/MGJXzG6tAK
— KEMRI-Wellcome Trust (@KEMRI_Wellcome) November 14, 2022
The participants were very engaged with the course, very keen to learn as much as possible and were not afraid to ask questions. This made the course more interactive and led to some lengthy discussions and exchanges of ideas after the lectures, which we hope will have helped the participants understand MR better and created networking opportunities. It also made it more enjoyable for the tutors to deliver the course!
We received many positive comments:
“Thanks for putting together a great course! We hope there will more followup courses and access to faculty to collaborate”
“This was a great course! I got enough skill and know how to appreciate what MR is, how to conduct an MR study, critically appraise one, and write a scientifically sound MR article. Many thanks to the organizers and the tutors!”.
Data gaps in Africa
The course also highlighted some challenges. The biggest was the lack of data from individuals of continental African ancestry. Even though summary statistics from many large genome-wide association studies are publicly available, and there is a movement to diversify genetic data collection and analysis, most participants in these existing studies are of European ancestry.
In addition, the African data that is available tends to be from African American participants, which may not be representative of the continental African participants nor individuals within African countries (e.g., due to admixture). For example, inter-marriage between African tribes rarely happens, causing each of the tribes within African countries to be genetically unique. Consequently, the results from MR analyses using data from African Americans are likely not be applicable to continental African populations. This huge diversity in African populations also means that many of the available reference panels and datasets that include some African individuals are likely not representative of many individuals on the continent. Furthermore, genotyping chips were designed and tested on European population and therefore current genotyping may not target all regions of interest to African populations.
Some of the feedback reflected this challenge:
“It was Excellent, Can we give examples on the African settings”
“It was a good introduction to MR for molecular traits. More examples in infectious diseases that affect Africans would be useful additions”.
What was abundantly clear was that more funding needs to be directed to African epidemiologists and genetic epidemiologists to bolster the availability of data necessary for causal analyses on the continent, and more international collaborations are required to support such funding efforts. Therefore, as researchers, we need to do more to encourage data collection from the African continent and bring more awareness to the unique challenges associated with studying populations of African ancestry.
Day 3 of the #RanMen conference and the team going deeper into assessing Mendelian Randomization assumptions. #GEfAS pic.twitter.com/HmrjT1cxNE
— KEMRI-Wellcome Trust (@KEMRI_Wellcome) November 16, 2022
Many of the course participants showed an interest in collaborating with us and the wider IEU after the course. To give some examples of collaborations this course has facilitated thus far (with the hope of many more!), Dr. John Muriuki will be coming to IEU to collaborate next year to discuss his work at the Kemri Wellcome Trust centre and Jarra Manneh (from MRC Unit in The Gambia) has applied for Wellcome Trust PhD funding to study the relationship between climate change and epigenetics with Prof. Caroline Relton within the IEU. In addition, Dr. Oscar Nyangiri (from Makerere University, Uganda) has plasma samples from children and adults affected by schistosomiasis and is interested in using these samples to obtain proteomic data to complement their genomic data. He and Dr. Lucy Goudswaard have met with the commercial proteomics companies Olink and SomaLogic to determine feasibility. Dr. Kaitlin Wade and Dr. Oscar Nyangiri have also discussed whether MR could be applied to understand whether schistosomiasis infection could be a causal risk factor for health and disease. To further this research question, he is seeking advice on how to obtain additional funding or seeking groups who may have a shared interest in this area.
Due to the popular demand and success of running this course, Dr. Sarah Atkinson and Dr. Chin Yang Shapland will collaborate and seek funding to establish additional training course for MR at the new Training centre at Kemri Wellcome Trust. And we have been in touch with other course participants and look forward to forming more collaborations. Please get in touch if you would like to learn about the work being undertaken by any of these brilliant scientists or have in an interest in collaborating with them, which we would be delighted to facilitate.
It was an incredible one week of genetic epidemiology and causal inference training hosted by @Sarah_AtkinsonS and facilitated by a brilliant team of instructors from @mrc_ieu. Lots of potential for collaboration that could lead to more #GWAS from continental Africa. #GEfAS pic.twitter.com/3T5DaAqAnU
— Jarra (@JarraJmanneh) November 19, 2022