Back

Adapting Epigenetic Clocks for Cell-Free DNA High-Throughput Sequencing Data

Li, G.; Huang, W.; Zhao, X.; Wu, J.; Guo, Y.; Chen, L.; Cao, X.; Yang, Z.; Jiang, S.; Hu, B.; Wang, Y.; Tan, D.; Tong, V.; Tang, C.; Feng, X.; Hu, X.; Ouyang, C.; Zhou, G.

2025-12-01 bioinformatics
10.1101/2025.11.27.690895 bioRxiv
Show abstract

Cell-free DNA (cfDNA) methylation sequencing holds promise for developing epigenetic aging clocks. However, current clocks--primarily trained on array-based data--do not readily generalize to high-throughput sequencing (HTS) cfDNA profiles. Using datasets with technical replicates encompassing HTS data from both cfDNA and gDNA, alongside gDNA methylation array data, we systematically assessed factors influencing clock accuracy and reproducibility. We identified key strategies to overcome HTS-specific challenges: maintaining [≥]10x mean target depth, applying elastic net regression with strong L2 regularization, and imputing unreliable beta-values. Further, transfer learning more effectively corrected platform biases than DNA type-related biases, and it enhanced performance robustly across multiple independent cohorts in aging and disease-related applications. Our findings demonstrate that array-derived epigenetic clocks can be effectively adapted to cfDNA sequencing data. This work offers critical methodological insights and practical guidelines, advancing the feasibility of minimally invasive aging assessment using cfDNA.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

1
Nature Communications
○ 5641 papers in training set
Top 16%
11.8%
Similar papers in this journal
2
Briefings in Bioinformatics
◐ 354 papers in training set
Top 0.5%
10.9%
Similar papers in this journal
3
Nature Aging
○ 60 papers in training set
Top 0.2%
9.6%
Similar papers in this journal
4
Genome Biology
○ 637 papers in training set
Top 2%
6.2%
Similar papers in this journal
5
Advanced Science
○ 286 papers in training set
Top 1%
5.1%
Similar papers in this journal
6
eLife
● 5828 papers in training set
Top 26%
4.8%
7
Genome Medicine
○ 183 papers in training set
Top 1.0%
4.0%
50% of probability mass above
8
Communications Biology
○ 993 papers in training set
Top 4%
3.5%
9
Scientific Reports
○ 3612 papers in training set
Top 29%
3.4%
10
iScience
○ 1154 papers in training set
Top 7%
3.2%
11
Cell Genomics
○ 172 papers in training set
Top 2%
2.6%
12
Science Advances
● 1243 papers in training set
Top 14%
2.4%
13
Epigenetics
○ 50 papers in training set
Top 0.2%
2.4%
14
Genome Research
● 468 papers in training set
Top 3%
2.4%
15
Nucleic Acids Research
◐ 1281 papers in training set
Top 8%
2.1%
16
Bioinformatics
◐ 1204 papers in training set
Top 6%
1.9%
17
GeroScience
○ 109 papers in training set
Top 1%
1.7%
18
Computational and Structural Biotechnology Journal
● 242 papers in training set
Top 5%
1.1%
19
Genomics, Proteomics & Bioinformatics
188 papers in training set
Top 2%
1.1%
20
Nature Methods
○ 385 papers in training set
Top 5%
1.1%
21
PLOS Computational Biology
● 1863 papers in training set
Top 18%
1.0%
22
BMC Bioinformatics
○ 457 papers in training set
Top 6%
0.8%
23
PLOS ONE
● 5266 papers in training set
Top 59%
0.8%
24
Aging
○ 75 papers in training set
Top 2%
0.8%
25
Aging Cell
○ 165 papers in training set
Top 2%
0.8%
26
Cell Reports
○ 1498 papers in training set
Top 29%
0.6%
27
eBioMedicine
○ 183 papers in training set
Top 8%
0.6%
28
GigaScience
◐ 212 papers in training set
Top 5%
0.6%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.