Page 139 - SAMRC AnnualReport 2025-26

P E R FOR M A NC E I N FOR M A T I ON

Postgraduate training remains central to the specific therapeutic vulnerabilities in key cancer
Platform’s capacity-building mandate, with a strong drivers such as TP53 and KRAS. This supports
emphasis on supporting historically disadvantaged precision oncology and enhances the identification
individuals. MSc and PhD students are actively of clinically relevant treatment targets.
engaged in research areas such as AI-driven
genomics, infectious disease surveillance, and rare Ongoing work also focuses on evaluating model
disease research. performance using African genomic datasets to
reduce bias and improve the generalisability of AI
AI in Medical and Health Research tools across diverse populations.

The Genomics Platform is actively integrating artificial In rare disease research, DDD-Africa is leveraging
intelligence (AI) to advance health research, disease AI tools such as Face2Gene through the DSI-Africa
surveillance, and data-driven decision-making. grant to improve diagnostic outcomes in settings
with limited access to clinical genetic expertise.
In collaboration with the Biomedical Research
and Innovation Platform and the Environmental Complementary initiatives are combining long-
Health Research Unit, the Platform contributes to read Oxford Nanopore sequencing with AI-based
the WSARP Programme, which monitors SARS- analytical approaches to investigate unsolved
CoV-2 circulation through wastewater surveillance. developmental disorder trios, with a focus on
Building on this foundation, AI-driven predictive structural variation, complex genomic regions,
models have been developed that integrate and episignatures.
wastewater data with clinical case information to In addition, large language models are being
analyse trends and forecast infection dynamics, applied to optimise bioinformatics workflows,
thereby strengthening early-warning systems and while tools such as Franklin (Genoox) support
supporting proactive public health responses.
variant interpretation. The DeepFASTQ framework
In genomics, machine learning is applied to improve further enhances sequencing data quality by
the interpretation of cancer-associated variants. applying genomic language models to FASTQ files,
By integrating mutation profiles, gene expression improving error detection, reducing turnaround
data, CRISPR gene-dependency screens, and drug times, and strengthening the reliability of genomic
response datasets, the Platform identifies mutation- analysis in South Africa.

The SAMRC is driving genomic sovereignty in Africa through its Genomics Platform and partnerships with
projects like the AfricaBP.

SA M R C A N N U A L R EP O R T 2 0 2 5 / 2 0 2 6 137