Scholarlink

Background <p>Cell-free DNA (cfDNA) coverage patterns are emerging as potential low-invasive approaches to determine transcription factor (TF) activation status in tumors; however, the accuracy of this approach has been confirmed for only a few TFs. Here, using paired tumor/blood samples collected from two xenograft tumor models of human liver cancer (HepG2, gain-of-function mutation in <i>CTNNB1</i>; HuH7, wild type), we compared the accuracy of inferences of TF activity made using tumor-derived ATAC-seq data and plasma-derived cfDNA whole-genome sequencing (WGS) data.</p> Results <p>ATAC-seq and cfDNA-WGS data from the HepG2 model both showed higher activation of two downstream target TFs of CTNNB1 in the Wnt signaling pathway compared with the HuH7 model. Expanding the study to 377 TFs, each with 10,000 TF-binding sites, revealed significant concordance between the data sets for both models (Spearman’s <i>ρ</i>: HepG2, − 0.90; HuH7, − 0.85). To elucidate potential factors that may influence the accuracy of cfDNA-based inferences of TF activity in the clinic, we prepared simulation samples by mixing tumor-derived cfDNA with healthy-donor cfDNA; we found that a minimum of 5 × sequencing coverage and a tumor fraction of at least 3% were required to obtain accurate inferences of tumor type–specific TF activity.</p> Conclusions <p>cfDNA-based TF activity inference is applicable to more than 370 TFs and can be used to accurately estimate tumor-specific TF activities. We believe that this study supports a fundamental principle that will be useful for future investigations of the potential of cfDNA analysis.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Cell-free DNA–based inference of the activities of 370 + transcription factors mirrors their activities in tumors

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月7日

查看全文
  • Ryuji Tamaki,
  • Koji Sagane,
  • Shuyu Dan Li,
  • Taisuke Hoshi

摘要

Background

Cell-free DNA (cfDNA) coverage patterns are emerging as potential low-invasive approaches to determine transcription factor (TF) activation status in tumors; however, the accuracy of this approach has been confirmed for only a few TFs. Here, using paired tumor/blood samples collected from two xenograft tumor models of human liver cancer (HepG2, gain-of-function mutation in CTNNB1; HuH7, wild type), we compared the accuracy of inferences of TF activity made using tumor-derived ATAC-seq data and plasma-derived cfDNA whole-genome sequencing (WGS) data.

Results

ATAC-seq and cfDNA-WGS data from the HepG2 model both showed higher activation of two downstream target TFs of CTNNB1 in the Wnt signaling pathway compared with the HuH7 model. Expanding the study to 377 TFs, each with 10,000 TF-binding sites, revealed significant concordance between the data sets for both models (Spearman’s ρ: HepG2, − 0.90; HuH7, − 0.85). To elucidate potential factors that may influence the accuracy of cfDNA-based inferences of TF activity in the clinic, we prepared simulation samples by mixing tumor-derived cfDNA with healthy-donor cfDNA; we found that a minimum of 5 × sequencing coverage and a tumor fraction of at least 3% were required to obtain accurate inferences of tumor type–specific TF activity.

Conclusions

cfDNA-based TF activity inference is applicable to more than 370 TFs and can be used to accurately estimate tumor-specific TF activities. We believe that this study supports a fundamental principle that will be useful for future investigations of the potential of cfDNA analysis.