UITOTO: a software for generating molecular diagnoses for species descriptions

Area of Science:

• Bioinformatics
• Taxonomy
• Molecular Biology

Background:

• Millions of undescribed species require molecular diagnoses for identification.
• Existing software lacks capabilities for deriving and validating state-specific, contrastive molecular diagnostic combinations (DMCs).

Purpose of the Study:

• Introduce UITOTO, a novel software for identifying, testing, and visualizing DMCs.
• Evaluate UITOTO's performance against existing tools using empirical datasets.

Main Methods:

• UITOTO employs a weighted random sampling algorithm (Jaccard Index) to build candidate DMCs.
• A validation module tests DMCs against large sequence databases for contrastiveness.
• Performance was assessed using three empirical datasets from Diptera and Lepidoptera.

Main Results:

• UITOTO successfully identifies, tests, and visualizes DMCs, addressing limitations of current software.
• UITOTO's DMCs demonstrated superior performance based on classification metrics (e.g., F1 Score) compared to other software.
• The study provides guidelines for generating molecular diagnoses and a user-friendly GUI.

Conclusions:

• UITOTO offers a robust solution for generating reliable molecular diagnoses, crucial for species description.
• Balancing specificity and length is key for effective molecular diagnoses, similar to morphological approaches.
• The software facilitates the process of species identification and taxonomic research.