Ability of methylation sequencing to enhance interpretation of clonal hematopoiesis
Clonal hematopoiesis of indeterminate potential is considered the presence of myeloid neoplasm-associated somatic mutations in otherwise healthy people without underlying myeloid neoplasia or unexplained blood count abnormalities. Some clinicians periodically monitor the size of the clonal population by measuring variant allele fraction (VAF) in sequencing assays. The proportions and types of clonal hematopoiesis of indeterminate potential (CHIP)-bearing cells (granulocytes, monocytes, lymphocytes) can vary in response to factors such as infection and inflammation, which may confound clonal expansion interpretation based solely on VAF. The authors conducted a study in which they described a novel, cost-effective, targeted DNA methylation sequencing assay to infer cell-type proportions and assess clonal expansions. The study focused on 91 samples from 34 patients previously assessed for CHIP. The authors used New England Biolabs’ commercially available enzymatic methyl-seq (EM-seq) assay, which detects 5-methylcytosine (5mC) and 5-hydroxymethylcytosine (5hmC). Cell-type deconvolution was performed with the R package methylCC, a statistical method that uses a latent variable model to estimate cell composition in a methylation platform-agnostic manner. In this model, regions of the genome in which each cell type is either clearly methylated or unmethylated are identified and then modeled as latent states. These latent states are biologically driven and, therefore, methylation-method independent. Linear modeling was used in the study to identify relationships between changes in CHIP VAFs and changes in methylation-predicted cell-type proportions. The authors found that linear relationships between these changes could predict the VAF trajectory (expanding, shrinking, or no change) of CHIP clones with a high degree of accuracy. Using cell-type proportion-informed VAF interpretation, CHIP clonal trajectory was reclassified in 57.1 percent of paired samples (16 of 28 pairs) compared with VAF-only classification. A limitation of this technique was the low correlation of methylation-predicted monocyte proportions and CBC-determined monocyte proportions. Granulocytes were also found to be consistently underrepresented in methylation-based predictions. Despite these limitations, the authors concluded that integrating methylation-based cell-type proportion data with VAF measurements significantly enhanced the interpretation of clonal trajectories in hematopoietic cells. This method potentially could be used to evaluate patients with clonal cytopenias of undetermined significance, a well-studied population with conferred risk of progression to hematological cancers.
Parker AC, Van Amburg JC, Heimlich JB, et al. Methylation sequencing enhances interpretation of clonal hematopoiesis dynamics. Blood. 2025;145(9):988–992.
Correspondence: Dr. Alexander G. Bick at alexander.bick@vumc.org