Automated landmarking via multiple templates

Area of Science:

• Morphometrics
• Computational Biology
• Evolutionary Biology

Background:

• Manual landmarking is time-consuming and prone to errors.
• Existing automated methods struggle with high sample variability due to single-template bias.

Purpose of the Study:

• Introduce MALPACA, a fast, open-source pipeline for automated landmarking using multiple templates.
• Present a K-means method for template selection to improve MALPACA's performance.
• Enhance accuracy and efficiency in quantifying complex morphological phenotypes.

Main Methods:

• Developed MALPACA (Multi-template Automated Landmarking Pipeline for Comparative Analysis).
• Implemented a K-means clustering algorithm for optimal template selection.
• Validated performance on single and multi-species datasets.

Main Results:

• MALPACA significantly outperforms single-template automated landmarking methods.
• K-means template selection is more effective than random selection.
• The pipeline demonstrates efficiency and reproducibility for large-scale morphological variation.

Conclusions:

• MALPACA offers an efficient and reproducible solution for landmarking highly variable samples.
• The multi-template approach accommodates significant morphological diversity, crucial for evolutionary studies.
• Open-source software is provided to support the research community.