Understanding all the genes active in a complex tissue can be difficult because of technical hurdles like assembly mistakes, many versions of the same gene, and problems with how sequences are labeled. These challenges are especially true in venom glands, where many different toxin genes are expressed at varying levels.

Researchers at Shahid Chamran University of Ahvaz developed a new computational approach called ToxIR to tackle these challenges for scorpion venom. ToxIR is an RNA sequencing pipeline that puts together several tools in a smart way to build and interpret the full set of toxin transcripts accurately, even the ones that are rare or look very similar to others.

The pipeline starts with rigorous quality checks and cleaning of raw sequencing data, then uses a specialized assembly program to construct full-length transcripts without relying on an existing reference. After assembly, ToxIR uses carefully curated databases and structural prediction tools to spot features typical of toxins, like signal peptides and specific domain patterns. All of this helps separate real toxin transcripts from noise and errors.

When tested on the venom glands of the scorpion Odontobuthus doriae, ToxIR identified 378 possible toxin transcripts, including 192 high-confidence ones and 23 completely new toxins that differ significantly from known sequences. These toxins target sodium, potassium, and chloride channels, key players in nerve and muscle function.

By improving how venom transcriptomes are analyzed, ToxIR makes it easier to discover new bioactive peptides that could have medical or research applications. Because the pipeline is flexible, it can be applied to many venomous species and help scientists explore toxins with greater accuracy than before.

Availability – All scripts required to run the ToxIR pipeline, along with detailed step-by-step usage instructions, are provided in the README file available at: https://doi.org/10.5281/zenodo.17931954