Introduction
Monitoring genomic diversity is essential in aquaculture to maintain long-term population health, productivity, and ensure the sustainability of the system, reducing undesired and uncontrolled declines in performance and eventually economic loss (Houston et al. 2020). In this context, current high-throughput genomic tools have the potential to address this challenge, by tailoring individual-based strategy. However, these approaches rely on individual fish sampling, a procedure that is not sustainable for all marine finfish farms, due to high costs, logistical constraints, and animal handling stress. To address these challenges, environmental DNA (eDNA) analysis represents a promising non-invasive approach for assessing biodiversity and its use is expanding in several fields (Yu et al., 2026). However, despite its traditional use for species detection and small-scale genetic analyses, its application to recover population genome-wide information remains unexplored. In this study, we evaluated the feasibility of untargeted whole-eDNA sequencing to estimate broodstock genetic diversity in gilthead seabream (Sparus aurata) under commercial production conditions, by comparing genome-wide diversity estimates obtained from eDNA with those derived from tissue-based DNA pools (Pool-seq).
Materials and Methods
A total of four recirculating aquaculture system tanks each containing from 57 to 67 broodstock fish were used for this study. eDNA from these tanks was collected by filtering two to three liters of water using filters with various pore sizes (0.2, 1.2, and 5.0 μm) and DNA was then extraction with an in house cetyltrimethylammonium bromide (CTAB) based protocol. Fin clips of the same individuals were collected, the DNA was individually extracted with a similar CTAB based protocol, and equimolar pools were assembled. eDNA and tissue derived DNA pools (DNA pool) were sequenced (paired end whole genome sequencing) and sequenced reads were mapped to the gilthead seabream reference genome using standard pipelines. After quality filtering, variant calling was performed on high-quality reads, retaining biallelic single nucleotide polymorphisms (SNPs) based on stringent quality and depth thresholds. These SNPs were used to estimate heterozygosity and genetic differentiation, and to compare eDNA- and tissue-derived genome-wide diversity estimates using ΔFIS-like and FST metrics.
Results
High-throughput sequencing generated an average of ~744 million reads for DNA pools and ~262 million reads for eDNA samples. While >99% of reads from DNA pools mapped to the gilthead seabream genome, eDNA samples showed substantially lower mapping rates (up to ~34–51%, depending on filter pore size), reflecting the presence of non-target DNA. Sequencing depth and SNP recovery in eDNA samples were markedly lower and more variable compared to DNA pools (~120× depth and >6M SNPs). Among filters, performance increased with pore size: 0.2 μm filters showed the lowest depth (2.7–5.5×) and SNP recovery (3k–19k SNPs), while 1.2 μm filters improved both metrics (3.0–8.0×; 4k–875k SNPs). The 5.0 μm filters achieved the highest depth (3.3–12.9×) and SNP counts (17k–4.1M), approaching the genomic resolution of DNA pools. This increase in sequencing performance was consistent with improved concordance in genetic diversity estimates (lower FST and ΔFIS closer to zero), as reported in Table 1.
Filter (μm)
Depth (×)
SNPs (n)
ΔFIS range
FST range
0.2
2.7–5.5
3k–19k
~0.01
0.03–0.32
1.2
3.0–8.0
4k–875k
~0.01
0.01–0.30
5
3.3–12.9
17k–4.1M
0.0017–0.01
0.01–0.04
Table 1. Sequencing performance and genome-wide diversity estimates from eDNA across filter sizes, compared with DNA pools (depth, SNPs, ΔFIS, FST).
Discussion
These results show that whole eDNA sequencing can recover genome-wide patterns of genetic diversity comparable to those obtained from DNA pools. Despite lower sequencing depth and the presence of non-target DNA, eDNA samples produced consistent estimates of heterozygosity and genetic differentiation across tanks. The improvement observed with increasing filter pore size suggests that larger filters are more effective in capturing host-derived DNA, likely through the retention of intact cells and cellular debris rather than fragmented extracellular DNA. This is reflected in the higher sequencing depth, greater SNP recovery, and improved concordance with DNA pools observed in the 5.0 μm fraction. Overall, these findings support the feasibility of using eDNA as a non-invasive approach for genome-wide monitoring of broodstock genetic diversity at the population level in aquaculture systems and may offer a practical alternative to individual-based sampling in population-level genetic monitoring in routine broodstock management.
References
Houston, R.D., Bean, T.P., Macqueen, D.J., Gundappa, M.K., Jin, Y.H., Jenkins, T.L., Selly, S.L.C., Martin, S.A.M., Stevens, J.R., Santos, E.M., Davie, A., Robledo, D., 2020. Harnessing genomics to fast-track genetic improvement in aquaculture. Nat. Rev. Genet. 21, 389-409. https://doi.org/10.1038/s41576-020-0227-y
Yu, Y., Shao, M., Zhou, H., Su, F., Guan, P., Ding, L., Ma, T., Zhou, B., Xu S, Song S, Qu B., 2026. Comparison between eDNA and traditional morphological methods for fish diversity monitoring in rivers. Sci Rep. 16:3098. doi: 10.1038/s41598-025-32964-1