A new paper co-authored by Illinois and African scientists describes how an African bioinformatics network tested researchers’ ability to perform computational genomics tasks that will aid in studying disease.

Since 2010, the Human Heredity and Health in Africa (H3Africa) initiative has taken a genomics-centered approach to studying disease susceptibility in Africa.

One of the initiative’s priorities is to train African scientists and reduce their dependency on collaborators from developed countries when pursuing research.

However, most of H3Africa’s research projects have relied on collaborators from developed countries for both the generation and analysis of data.

To help develop the bioinformatics capacity of African scientists and support H3Africa’s mission, a program known as H3Africa Bioinformatics Network (H3ABioNet) was created.

At H3ABioNet’s first general assembly, Victor Jongeneel, former Director of Bioinformatics at the IGB and a key participant in H3ABioNet, proposed an idea: to create a set of exercises that could assess H3Africa’s research groups’ abilities in managing and analyzing genomic datasets.

Once the exercises were created, three groups from H3ABioNet volunteered to participate. They were required to complete data-analysis tasks — such as analyzing genotyping data from genome-wide association studies — that are commonly faced by bioinformatics facilities.

The exercise involved three steps: preparation of the teams and setting up the needed software, receiving and analyzing test data sets and writing a report, and an evaluation of these reports by an expert panel.

A set of “standard operating procedures,” or SOPs, was also given to the groups to guide them in their analyses. The High-Performance Biological Computing (HPCBio) group at the IGB, National Center for Supercomputing Applications (NCSA), and the Roy J. Carver Biotechnology Center worked with groups in South Africa to create these procedures and prepare test datasets for the exercise.

The results of the exercise were presented in a paper published in PLOS Computational Biology (DOI: 10.1371/journal.pcbi.1005419).

Overall, the implementation of the exercise was considered a success, as the participating groups were able to improve their own operations and provide feedback that could be used for following exercises.

“This is a continuing process, which is going to continue into the second funding period for H3Africa and H3ABioNet,” Jongeneel said. “We have actually had at least half a dozen nodes taking one of the exercises in the first half of 2017. The SOPs and datasets are being constantly revised and brought up to date.”

However, the paper described the level of participation in the exercise as “extremely disappointing,” as only three of the 32 groups in H3ABioNet chose to participate.

“We were very pleased that those nodes who took the exercises did very well, and of course we were disappointed that not more of them participated,” Jongeneel said.

Time and effort, inadequate staffing, and inexperience with certain skills were all given as reasons why groups chose not to participate — but the paper stated a fear of failure was an additional, unspoken reason.

“The lack of self-confidence of many African scientists has been a factor in the symmetry of their relationships with first world collaborations and, in our opinion, is one of the issues to be addressed in capacity-development efforts,” the paper explained.

H3ABioNet plans to encourage more research groups to participate in future exercises by engaging with researchers and letting them know the benefits of participating. This effort has already paid off, as eight more groups have participated in early 2017.

“A key mission of H3Africa is to develop the capacity of African groups to perform genomic research independently, or in collaboration with other African researchers,” Jongeneel said. “Our work is helping make this possible by setting objective and measurable criteria for their capacity to analyze complex data.”

This research, which was funded by the National Human Genome Research Institute and the Office of The Director, National Institutes of Health, gave participating groups the chance to validate their ability to pursue genomic research, according to Jongeneel.

“Having successfully submitted to one or more of the exercises will help the H3ABioNet nodes establish credibility as collaborators for H3Africa research projects,” he said. “This will help the H3Africa project overall by reducing the need for collaborators in developed countries, who often take the scientific spotlight away from the African scientists.”