Short Tree, Long Tree, Right Tree, Wrong Tree: New Acquisition Bias Corrections for Inferring SNP Phylogenies.

Area of Science:

• Genomics
• Evolutionary Biology
• Bioinformatics

Background:

• Single nucleotide polymorphisms (SNPs) are valuable genomic markers for phylogenetic studies due to their abundance and ease of collection.
• Restriction site associated DNA sequencing (RADseq) is a popular method for collecting SNP data, but best practices for its phylogenetic application require further investigation.

Purpose of the Study:

• To assess the accuracy of RAD loci for phylogenetic inference using computer simulations and empirical double digest RADseq (ddRADseq) data from the lizard family Phrynosomatidae.
• To compare phylogenetic analysis using full RAD sequences versus SNP data alone.
• To introduce and evaluate new methods for correcting acquisition bias in SNP-only phylogenetic analyses.

Main Methods:

• Computer simulations were used to test phylogenetic accuracy under different data partitioning strategies.
• Double digest RADseq (ddRADseq) data were generated for the lizard family Phrynosomatidae.
• Phylogenetic analyses compared the use of full RAD loci (SNPs and invariant sites) versus SNP data only.
• Two novel acquisition bias correction methods were developed: a conditional likelihood method and a reconstituted DNA approach.

Main Results:

• Using full RAD sequences, including invariant sites, resulted in superior branch length and topological accuracy compared to using SNPs alone.
• Both conditional likelihood and reconstituted DNA bias correction methods showed improved branch length accuracy, with the reconstituted DNA approach performing better, especially with increased missing data.
• Phylogenetic analyses of empirical Phrynosomatidae data using concatenation and coalescent-based species tree methods supported established relationships, indicating RAD loci contain robust phylogenetic signal.

Conclusions:

• Phylogenetic analyses using full RAD loci sequences are preferable to SNP-only analyses for improved accuracy.
• The reconstituted DNA approach offers a promising method for correcting bias in SNP-only RADseq phylogenetics.
• Careful consideration of model assumptions is crucial for accurate phylogenetic inference from RAD loci, particularly when branch lengths are important.