PeptideMiner, a new search tool, discovered 10 novel natriuretic peptides and 57 novel insulin-like sequences from animal venoms, with some showing activity at human insulin receptors.
57 novel insulin-like peptidesDiscovered from marine cone snail venoms, with confirmed activity at human insulin receptors
What the researchers found
PeptideMiner uses profile-hidden Markov models to discover peptide families across species despite high sequence divergence. Benchmarking showed it outperformed existing methods.
Applied to venom transcriptomes (including 24 previously unpublished datasets), PeptideMiner identified 10 novel natriuretic peptides from distantly related species and 57 novel insulin-like sequences from marine cone snails. Chemical synthesis and testing of newly identified conoinsulins at human insulin receptors confirmed biological activity, validating the approach for discovering pharmacologically relevant peptides.
Why it matters
Only a fraction of the estimated neuropeptide diversity has been functionally characterized. PeptideMiner accelerates discovery by finding peptides that other tools miss due to sequence divergence. The cone snail insulin-like peptides are particularly exciting because nature has already evolved molecules that interact with human drug targets.
How the study worked
The researchers built PeptideMiner using profile-hidden Markov models trained on known neuropeptide families. They validated it against existing methods using benchmark datasets, then applied it to venom transcriptome databases including 24 new datasets. Newly discovered insulin-like peptides from cone snails were chemically synthesized and tested for activity at human insulin receptors.
What this study cannot tell us
While PeptideMiner outperformed existing methods, no computational tool can find all peptides — novel families with no known relatives will still be missed. The functional characterization was limited to insulin receptor binding; in vivo efficacy, toxicity, and pharmacokinetics of the new peptides were not assessed. The tool depends on available transcriptome and genome data quality.
How to read the evidence
This is a computational tool development and validation study with experimental confirmation through chemical synthesis and receptor binding assays. The methodology is rigorous and the tool is benchmarked, but the discovered peptides are at the earliest stage of pharmacological characterization.
When this study was published
Published in 2025 in GigaScience, this is a very recent contribution to the rapidly growing field of computational peptide discovery.
The bigger picture
Nature is an enormous untapped pharmacy of bioactive peptides. Cone snail venoms alone have yielded one FDA-approved drug (ziconotide for pain). Tools like PeptideMiner could dramatically accelerate the discovery of peptide-based drug leads from the vast diversity of animal peptidomes, particularly from understudied species.
Questions still open
- Could any of the 57 novel conoinsulins be developed into fast-acting insulin analogues for diabetes treatment?
- What other drug-relevant peptide families might PeptideMiner uncover from unexplored animal venoms?
- Can PeptideMiner be applied to metagenomics data to discover peptides from unculturable organisms?
Common questions
Why are cone snail venoms useful for drug discovery?
What are conoinsulins and why are they interesting?
Read the original research
PeptideMiner-neuropeptide discovery across the animal kingdom.
GigaScience, 14
Citation
Mendel, Helen C; Hopping, Gene; Undheim, Eivind A B; Zuegg, Johannes; Lewis, Richard J; Forbes, Briony E; Kaas, Quentin; Muttenthaler, Markus. (2025). PeptideMiner-neuropeptide discovery across the animal kingdom.. GigaScience, 14. https://doi.org/10.1093/gigascience/giaf078