Back

Eco-genomic analysis uncovers precision-conservation targets for the western Pacific's southernmost salmonid

Lee, Y.-C.; Shen, Z.-Y.; Lin, W.-R.; Wang, T.-Y.; Liu, M.; Yeh, Y.-C.; Chan, S.-F.; Lu, M.-Y. J.; Wang, H.-Y.; Liao, L.-Y.; Li, W.-H.; Huang, J.-P.; Tsai, I. J.; Shen, S.-F.

2025-09-15 genomics
10.1101/2025.09.09.675276 bioRxiv
Show abstract

Understanding how isolated small populations persist and adapt in adverse environments is instrumental to evolutionary and conservation biology. We combine a chromosome-level genome assembly, population resequencing and forward-time simulations to reconstruct the history and viability of the Formosan landlocked salmon (Oncorhynchus formosanus), now restricted to a handful of high-mountain headwaters in Taiwan. We estimate that this lineage has diverged from Japanese masu salmon over one million years ago and has no detectable gene flow for [~]50,000 years. It has accumulated extensive chromosome fusions and expansions of cold-adaptation gene families, qualifying it as a new species rather than a subspecies of Japanese masu salmon. Whole-stream sampling reveals an overlooked Hehuan-Creek population that retains high heterozygosity and has gained unique alleles. Life-table simulations show that the Hehuan population has a notably lower extinction risk and can persist or even grow under low-to-moderate typhoon frequency, whereas Qijiawan-Creek population would decline precipitously under the same or higher frequency . These findings contradict the notion that peripheral populations are likely genetically depleted and support stream-specific "precision conservation" in place of broad, untargeted translocations that could erode local adaptation potential. Thus, our genomic-ecological analysis has uncovered hidden strong resilience in a critically endangered, climate-threatened salmonid lineage.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"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.