Following on emerging understanding of the diversification process, many recent workers have considered infraspecific taxa as valuable for formally recognizing incompletely speciated entities. The distinction between a species and an infraspecific taxon represents a fundamentally subjective weighting of evidence, yet this points further to the need for an evidential basis for these decisions. We explore these concepts in Heuchera longiflora (Saxifragaceae), which is morphologically variable and has a disjunct range across several physiographic provinces in the eastern U.S. using a tiered sampling approach and a combination of Sanger and next generation sequencing (NGS) techniques. We investigated 56 populations with seven markers to investigate population structure, and for a subset of 12 representative populations we sequenced 277 nuclear markers to characterize gene tree discord. Using a variety of methods to overcome high levels of gene conflict, we find evidence for the traditionally recognized taxa H. longiflora var. longiflora and H. longiflora var. aceroides in molecular and morphological data partitions. Our results comprise a case study on the use of multiple sources of data and analytical methods to delimit infraspecific taxa.
Alignments of phased Sanger data
Sample codes follow the main text. "a" and "b" refer to reconstructed alleles. Numbers preceded by the caret symbol refer to assignments to a priori groupings used by BPP.
phased_loci.zip
Alignments of unphased Sanger data
Sample codes follow the main text. Numbers preceded by the caret symbol refer to assignments to a priori groupings used by BPP.
unphased_loci.zip
Floral morphometric data
Sample codes give the herbarium (Index Herbariorum codes), collector, and collector number. Floral number is sequential where multiple flowers were available on a specimen. For area codes, refer to text. All units are in centimeters except Floral length/width and Zygomorphy, which are unitless ratios.
floral_dimensions_final.xlsx
Leaf landmark data
The landmark file is in TPS format. Sample codes give the herbarium (Index Herbariorum codes), collector, and collector number. For area codes in the classifier file, refer to text.
leaf_landmarks_final.zip
NGS_gene_alignments
Alignments for the 277 gene alignments generated from NGS data.
Best ML trees from NGS data
The best ML trees for the 277 NGS genes, used for ASTRAL-II analyses
allbest.tre.renamed.tre
Bootstrapped ML trees from NGS data
The bootstrapped ML trees for the 277 NGS genes, used for ASTRAL-II analyses.
bootstraps.zip
Allele table
The allele table used to assign population samples to putative species for ASTRAL-II.
allele_table_longiflora.txt
Concatenated Sanger data with gaps coded
The seven Sanger loci concatenated, with a final binary partition for gap characters.
all_longiflora_trimmed_concatenated_gaps.phy
Partition file for concatenated Sanger data with gaps coded
Partition file, with gene-wise partitions and a single separate partition for gap data.
partitionlongifloratest_edited_gaps.txt
RAxML best tree: 7 Sanger loci, unpartitioned
The best tree with with boostrap frequencies, using the full Sanger dataset without gaps or partitions. Population codes follow the text; note this is an unrooted tree.
RAxML_bipartitions.longiflora_unpartitioned_julian.tre.renamed.tre
RAxML best tree: 7 Sanger loci, partitioned
The best tree with with boostrap frequencies, using the full Sanger dataset without gaps but with seven partitions. Population codes follow the text; note this is an unrooted tree.
RAxML_bipartitions.longiflora_partitioned_julian.tre.renamed.tre
RAxML best tree: 7 Sanger loci, unpartitioned, reduced missing data
The best tree with with boostrap frequencies, using reduced Sanger dataset without gaps or partitions. Sites with higher missing data were removed as described in the text. Population codes follow the text; note this is an unrooted tree.
RAxML_bipartitions.all_longiflora_30_missing_unpartitioned.tre.renamed.tre
RAxML best tree: 7 Sanger loci, partitioned, reduced missing data
The best tree with with boostrap frequencies, using the reduced Sanger dataset without gaps but with seven partitions. Sites with higher missing data were removed as described in the text. Population codes follow the text; note this is an unrooted tree.
RAxML_bipartitions.all_longiflora_30_missing_partitioned.tre.renamed.tre
RAxML best tree: 7 Sanger loci, partitioned, gap coding
The best tree with with boostrap frequencies, using the full Sanger dataset with gaps and with eight partitions. Population codes follow the text; note this is an unrooted tree.
RAxML_bipartitions.longiflora_partitioned_gaps_julian.tre.renamed.tre
RAxML best tree: 7 Sanger loci, partitioned, gap coding
The best tree with with boostrap frequencies, using the reduced Sanger dataset with gaps and with eight partitions. Sites with higher missing data were removed as described in the text. Population codes follow the text; note this is an unrooted tree.
RAxML_bipartitions.longiflora_partitioned_gaps_julian.tre.renamed.tre