Related Experiment Video
Updated: Aug 5, 2026
Selecting Multiple Biomarker Subsets with Similarly Effective Binary Classification Performances
Published on: October 11, 2018
CHAF1A identified as a candidate biomarker and potential therapeutic target in Burkitt lymphoma through integrated
Wen-Yuan Lin1, Bei-Li Chen1, Zhi-Mei Wu1
1Department of Hematology, The First Affiliated Hospital of Guilin Medical University, Guilin, China.
Background:
Burkitt lymphoma (BL) is a highly aggressive malignancy with limited effective treatments due to toxicity/resistance. Identifying and prioritizing candidate biomarkers and potential therapeutic targets from complex transcriptomic data, ahead of functional validation, remains a major unmet need.
Methods:
We integrated differential gene expression analysis across BL vs. control cohorts (Gene Expression Omnibus [GEO] datasets GSE43677/GSE12453) with Random Forest machine learning to prioritize candidates. Validated top genes via immunohistochemistry in an independent cohort (n = 10 BL, n = 10 reactive lymphoid hyperplasia [RLH] controls), followed by immune cell infiltration analysis.
Results:
This approach identified four candidate genes; only Chromatin Assembly Factor 1 Subunit A (CHAF1A) showed profound protein-level overexpression in BL tumors vs. RLH. In a single-dataset CIBERSORT analysis, CHAF1A expression showed exploratory correlational associations with several immune-cell fractions, most strongly a positive association with M0 macrophages; these in silico associations are hypothesis-generating and do not by themselves establish a mechanistic role in the tumor immune microenvironment.
Conclusion:
CHAF1A is identified as a candidate biomarker associated with BL that, in exploratory single-dataset analysis, also correlates with immune-cell infiltration; these findings nominate it as a potential therapeutic target that warrants functional validation in future studies.
