H3ABioNet, the University of Notre Dame and IBM Research Africa are organizing a joint hackathon on a complex biomedical research question on antimalarial drug resistance to prepare preliminary datasets for a future open collaborative challenge on drug combination prediction. While malaria, a disease caused most commonly by the protozoan parasite Plasmodium falciparum, remains a major global health challenge, with a global estimate of slightly less than half a million deaths in 2015 [WHO World Malaria Report 2015], the African continent carries a disproportionately high burden of the disease (88% of the world cases and deaths). Nowadays, the World Health Organization (WHO) recommends artemisinin-based combination therapies for the treatment of malaria cases due to P. falciparum. However, emerging resistance of P. falciparum to artemisinins has been observed over recent years in South-East Asia, and the WHO has stated that “The emergence of P. falciparum resistance to artemisinin is an urgent public health concern, threatening the sustainability of the global effort to reduce the malaria burden in all endemic regions.”In this context, it is the endeavor of consortia like H3ABioNet and IBM Research Africa to foster major breakthroughs in scientific research on diseases having a major impact on the African continent, hence the motivation for this collaboration.
Participants should demonstrate confirmed expertise in one or several of the above-mentioned domains and that they will bring an added value in the data analysis process. This workshop is clearly results-oriented, and training during the workshop will be limited to topic-specific sessions to implement efficient cross-talking between the different subgroups of expertise that will make up the group of workshop participants. It is NOT designed to be a training course. Participants are expected to benefit from learning to integrate skills from statistics, computing and biology to seeking solutions to biological problems in a collaborative manner and within a limited span of time. We anticipate that participants will work openly with each other and will therefore have the chance to learn from each other and develop critical thinking and may develop close collaborations and that the experiences on collaboration and scientific insights obtained will be published at a later date with participants as authors.
The main objective of this workshop is to normalize, pre-analyse and curate the microarray data generated in order to prepare for the launch of the two-fold DREAM Challenge that will be open in the near future to the research community.
After the data have been processed, workshop participants will be tasked with developing baseline models for predicting drug sensitivity and assessing the viability of the challenge. The specific aim is to evaluate the predictability of the level of drug resistance of P. falciparum isolates exposed to dihydroartemisinin (DHA), the active form artemisinin, based on their transcriptional response measured at two time-points in red blood cell cultures. The genetic background information on the isolates will be provided to the participants. The genome-wide transcriptional profiles will come in the form of microarray data.
Developing a satisfactory solution to this complex prediction problem will require participants to make use of a mix of skills pertaining to various areas of the broad bioinformatics field: microarray data analysis, drug response (drug mechanism of action and drug resistance), pharmacogenomics, vector-host interactions, discrete modeling, machine learning, biostatistics and computer programming. Therefore, we seek to gather a few motivated participants originating from the H3ABioNet/H3Africa consortium for a five-day workshop.
Classroom applications
Participant applications
The Ferdig Laboratory at the University of Notre Dame (Indiana, USA) is generating transcriptional data sets from a range of the malaria parasite isolates exposed to artemisinin in culture. This data needs to be preprocessed and prepared for an open challenge that will be launched under the DREAM Challenges initiative within the next 2 years. The challenge will invite participation of the whole research community in developing computational models for malaria drug combinations. Therefore, H3ABioNet, the University of Notre Dame and IBM Research Africa are organizing a data analysis workshop to prepare data for the launch of part 1 of a two-fold future open challenge on drug resistance prediction from transcriptional response in the context of (1) single drug treatment and (2) multi-drug treatment. In this latter challenge, the scientific problem will involve the prediction of synergistic and/or antagonistic effects of drug combinations from transcriptional responses to individual drugs.
Some work has already been published on drug sensitivity prediction using genomics data sets, including results from a previous DREAM Challenge [Costello et al, Nat. Biotechnol. 2014] or [Liu et al, Sci. Rep. 2016]. As part of the workshop, we also plan to review the biological foundations and hypotheses for predicting drug combinations
