Benchmarking Orthogroup Inference Accuracy: Revisiting Orthobench

Area of Science:

• Bioinformatics
• Computational Biology
• Phylogenetics

Background:

• Orthobench is the standard benchmark for evaluating orthogroup inference methods.
• It comprises 70 expert-curated reference orthogroups (RefOGs) across Bilateria.
• Existing RefOGs present challenges for current orthogroup inference tools.

Purpose of the Study:

• To re-evaluate and update the Orthobench benchmark using improved phylogenetic delineation.
• To assess the accuracy of widely used orthogroup inference methods with revised RefOGs.
• To provide an open-source suite for future Orthobench development and application.

Main Methods:

• Leveraged advanced tree inference algorithms and computational resources.
• Conducted extensive phylogenetic delineation of 70 expert-curated RefOGs.
• Utilized revised RefOGs to benchmark orthogroup inference method accuracy.

Main Results:

• Phylogenetic revision altered the membership of 31 out of 70 RefOGs.
• 24 RefOGs underwent extensive membership changes, while 7 had minor modifications.
• The study provides an updated assessment of orthogroup inference method performance.

Conclusions:

• The revised Orthobench benchmark offers a more accurate resource for evaluating orthogroup inference.
• Updated RefOGs facilitate a more reliable assessment of computational methods.
• An open-source suite is provided to support ongoing development and use of Orthobench.