cfDNA-Based Inference of 300+ Transcription Factor Activities Mirrors Their Activities in Tumors

Investigating the open chromatin structure and transcription factor (TF) activities in cell-free DNA (cfDNA) provides a potentially less-invasive approach to understanding cancer biology. However, the accuracy of inferring tumor TF activities from cfDNA has been explored for only a few TFs. We conducted a comparative analysis of TF activities between cfDNA and the Assay for Transposase-Accessible Chromatin using sequencing (ATAC-seq) from tumors.Both tumor ATAC-seq and plasma-derived cfDNA whole-genome sequencing (WGS) from the HepG2 xenograft model (with an activating CTNNB1 mutation) showed higher TCF/LEF TF activities, the downstream targets of CTNNB1 in Wnt pathway, compared to HuH7 (CTNNB1 wild-type). Moreover, we found significant concordance in TF activities for 377 TFs with 10,000 TF binding site each (Spearmans rank correlation coefficient for HepG2 and HuH7 -0.90 and -0.85, respectively). To assess the potential utility of cfDNA TFBS analysis in clinical samples, we mixed tumor-derived cfDNA with healthy donor cfDNA. Our findings revealed that a minimum of 5x sequencing coverage is required to infer tumor type-specific TF activity in cfDNA containing 3 percent tumor-derived cfDNA, which is clinically relevant across various cancer types. Additionally, a direct comparison between cfDNA WGS and tumor ATAC-seq provided insights into the global chromatin accessibility of cfDNA, crucial for estimating TF activity.Our results suggest that cfDNA can accurately estimate tumor-specific TF activities because tumor chromatin status is reflected in cfDNA. This establishes a fundamental principle for future oncology clinical studies to investigate cancer biology by analyzing TF activities in cfDNA