105P - Implementation of technical improvements in cfMeDIP-Seq library preparation

Background

Cell-free methylated DNA immunoprecipitation sequencing (cfMeDIP-Seq) is a promising approach to assess non-invasive methylomics. Despite its cost-efficiency and accuracy, the setup and processing of cfMeDIP-Seq are technically challenging. The original protocol developed by Shen et al. in 2019 can now be partially automated, potentially improving reproducibility and yield.

Methods

We designed a pilot study to evaluate the performance of cfMeDIP-Seq under 4 library preparation conditions. We collected 4 K₂EDTA and 4 PAXgene ccfDNA tubes (BD Vacutainer™) from the same healthy donor. We processed K₂EDTA and PAXgene tubes within 2 and 6 hours of blood draw respectively, as previously described (Dameri et al., 2023). We used 10 ng of cfDNA for library preparation. Each condition underwent 2 parallel preparation procedures: one following the manual protocol described by Shen et al., and the other following a semi-automated version of the original protocol using the IP-Star® Compact instrument (Diagenode). We generated a total of 16 libraries, 4 for each condition, and sequenced them on Illumina NovaSeq 6000 device.

Results

The number of uniquely mapped reads was comparable (p-value = 0.8) between semi-automated and manually processed libraries, with an average of 19.9 MReads per sample. The percentage of duplicates was reduced (p-value = 0.0012) in semi-automated libraries (median = 32.88; IQR = 31.94-34.06) compared to manually prepared ones (median = 41.48; IQR = 35.61-42.53). Semi-automated libraries showed a longer mean insert size (255 bp) compared to manually processed libraries (243 bp) with a p-value < 0.01. Principal component analysis revealed no major differences between K₂EDTA and PAXgene samples processed with the semi-automated protocol. We observed no significant differences across the 4 conditions in terms of GC content (47%, sd = 0.49).

Conclusions

In the present pilot study, we demonstrated that semi-automated library preparation outperformed the fully manual protocol, regardless of tube type but provided that K2EDTA were processed shortly after collection. Partial automation of fluidic handling seems extremely relevant to increase yield and quality of cfMeDIP-Seq experiments.

Editorial acknowledgement

Clinical trial identification

Legal entity responsible for the study

The authors.

Funding

Associazione Italiana per la Ricerca sul Cancro (AIRC) and 5x1000 by Italian Ministry of Health.

Disclosure

All authors have declared no conflicts of interest.