Landmark-based morphometrics with fewer landmarks: Some examples for medical entomology

Area of Science:

• Geometric morphometrics
• Entomology
• Biodiversity and Taxonomy

Background:

• Geometric morphometrics using 2D landmarks aids in distinguishing morphologically similar or cryptic arthropod taxa, crucial for public health.
• The common assumption is that more landmarks capture more shape information, thus improving discriminatory power.

Purpose of the Study:

• To challenge the assumption that increasing landmark number improves shape discrimination in geometric morphometrics.
• To evaluate if smaller landmark subsets can achieve equal or superior discriminatory power compared to full landmark sets.
• To simplify entomological surveillance by optimizing landmark selection for morphometric identification.

Main Methods:

• Compared unsupervised classification accuracy scores between full landmark sets (10-22 points) and smaller subsets.
• Utilized published data from six insect families (Culicidae, Glossinidae, Muscidae, Psychodidae, Reduviidae, Tabanidae).
• Employed two landmark subset selection strategies: landmark contribution to shape distance and reclassification scores from random samples.

Main Results:

• Small subsets of landmarks can equal or outperform full landmark sets in shape discrimination, excluding size variation.
• Validated results by accounting for chance reclassifications to ensure reliability.
• Demonstrated that fewer landmarks can be sufficient for accurate species identification.

Conclusions:

• Optimizing landmark selection simplifies entomological surveillance, accelerating identification and improving standardization.
• Reduced landmark sets minimize noise from problematic landmarks, enhancing accuracy for morphologically similar, medically important taxa.
• This approach offers a valuable, resource-efficient alternative or complement to molecular methods in entomological studies.