Integrating multi-omics data reveals hidden risk in childhood leukemia
A recent study published in Communications Medicine shows how combining multiple types of molecular data can improve risk assessment in childhood leukemia.
B-cell precursor acute lymphoblastic leukemia (BCP-ALL) is the most common childhood cancer. Many patients respond well to treatment, but some relapse despite being classified as low risk based on genetic markers. The reasons are not fully understood, which makes it difficult to identify these patients early.
Looking beyond single data types
Over the past decade, researchers have generated large datasets covering genomics, epigenomics, and transcriptomics. These layers are usually analyzed separately.
“We had not yet attempted a comprehensive integration of these layers to see the full picture,” says SciLifeLab researcher Olga Krali, first author of the study.
In this study, the team analyzed data from 1,231 patients. They used Multi-Omics Factor Analysis (MOFA) to reduce the data into ten “cross-modal elements” (CMEs). Each element captures patterns shared across data types, linking genes, DNA methylation, and drug responses.
Potential for refining subgroups
The analysis points to a subgroup of patients who appear low risk using standard genetic classification, but show features linked to relapse. By combining drug response data with DNA methylation profiles, the researchers identified signals that would otherwise remain undetected.
The pattern was seen within the “high hyperdiploidy” subtype, which is usually linked to good prognosis. Even after accounting for known clinical factors, this subgroup remained at higher risk of relapse.
“This insight could help explain why certain patients respond differently to treatment and may help us identify those at risk of poor outcomes despite a seemingly ‘safe’ genetic profile,” Krali says.
Next steps
The team is currently developing an online resource where researchers can explore the dataset, which includes molecular and drug response data from more than 1,200 patients. Their aim is to make it easier to compare findings and build on the results.
