# Impact of Nanopore Flow Cell Types on DNA Methylation Detection

## TL;DR

This study compares how different Nanopore flow cells affect DNA methylation detection in kidney cancer samples.

## Contribution

The study reveals systematic errors in R9 flow cells and shows how R10 and batch-correction software improve methylation accuracy.

## Key findings

- R9 overestimates methylation at promoters and underestimates it in intronic and intergenic regions.
- R10 flow cells show higher accuracy in DNA methylation detection due to improved sequencing mechanics.
- Batch-correction software reduces systematic errors and improves data comparability between flow cells.

## Abstract

Third-generation sequencing technologies have revolutionized the study of epigenetic characteristics in human diseases, with Oxford Nanopore Technologies (ONT) at the forefront of long-read sequencing. ONT has made rapid improvements in flow cell designs, which greatly increased its sequencing accuracy but, at the same time, led to some projects utilizing different flow cell types, mainly R9 vs. R10, across samples. Whether and how the flow cell types affect genome-wide DNA methylation detection remains incompletely understood. Here, we used both flow cell types to analyze 6 human renal cell carcinoma (RCC) samples and compared the results. While there was a highly significant correlation between 5-methylcytosine (5mC) detected by R9 and R10 flow cells, we also observed substantial differences. R9 flow cells over-estimated 5mC levels at hypomethylated chromatin regions, mostly at promoters, while under-estimated 5mC at hypermethylated chromatin regions, enriched at intronic and intergenic regions. Such deviations in detection were likely caused by substantially lower sequencing accuracy of R9 flow cells, due to its mechanics, especially having problems sequencing homopolymeric DNA elements, such as CpG islands, leading to both higher false-positive and false-negative detections. Interestingly, such systematic errors were largely mitigated by batch-correction software, improving data comparability. In summary, our study reports superior performance of R10 flow cells, leading to much higher accuracy in base sequencing and DNA methylation detection.