Integrating lineage-specific and universal genomic probes illuminates phylogenetic relationships and molecular...

Obtaining large and well-resolved phylogenetic trees for neotropical clades is challenging, as many species inhabit remote regions, and sampling often relies on herbarium specimens with highly degraded DNA. Target capture provides an effective solution for recovering molecular data from fragmentary material. However, data processing using tools generally designed for diploid organisms and single-copy loci is also challenging, particularly when events such as genome duplication and hybridisation have shaped lineage evolution. We used dual-hybridisation to integrate Ochnaceae-specific and universal probes to reconstruct the phylogenetic relationships of Sauvagesieae, a pantropical clade with ca. 90 species mainly distributed in the northern Andes, the Brazilian Espinhaço Range, and the Amazon-Guyana region. We tested different filtering strategies involving missing data and paralogs to assess probable sources of tree discordance and topological uncertainty. We found no significant benefit in reducing tree discordance after removing entire genes due to the presence of paralogs or a high amount of missing data. Removing fragmentary sequences instead improved alignments and increased branch support of gene trees. By quantifying the proportion of single nucleotide polymorphisms (SNPs), analysing the distribution of the allele frequencies, and gene-tree quartet frequencies, we found signals of polyploidisation and hybridisation, which could reduce resolution at internal nodes, particularly in mountain clades. Our results underscore the importance of exploring the complexities of target-capture data, not only to improve phylogenetic resolution but also to understand the sources of phylogenetic conflict and the underlying molecular evolutionary processes.