Stenodraba (Brassicaceae) included a group of eight species distributed along the Andes of South-Central Argentina and Chile. All of its species were subsequently transferred to other genera and are currently treated in Pennellia (Tribe Halimolobeae) and Weberbauera (Tribe Thelypodieae). However, the phylogeny of Stenodraba and its tribal placement were never analyzed using molecular data. The lack of such studies, as well as paucity of herbarium collections suggesting that some species are vulnerable and/or endangered, prompted us to address the molecular phylogeny of Stenodraba. For this purpose, we generated comprehensive molecular phylogenies using nuclear (ITS) and plastid (trnL-F and trnH-psbA) data, and conducted different niche comparisons in the environmental and geographic space using climate data processed both by ordination and species distribution modelling (SDM) techniques. The results demonstrated that Stenodraba belongs to the South American tribe Eudemeae and is related to the genera Aschersoniodoxa, Brayopsis, Dactylocardamum, and Eudema. Stenodraba species formed two, strongly supported clades, and although molecular data did not recover monophyly of the genus, this hypothesis was not significantly rejected. The main clades were differentiated in their climatic niches (both in the environmental and geographical spaces), and niche overlap was greater within than between clades. Systematic implications, including a key distinguishing Stenodraba from the remaining genera of Eudemeae and a synopsis of its species, are also provided.

#### Appendix1

APPENDIX 1. New sequences generated for this study. Species, specimen voucher, and Genbank accession numbers (ITS, trnL-F, trnH-psbA). For sequences download from the GenBank see APPENDIX 2.

#### Appendix2

APPENDIX 2. Taxa and GenBank accession numbers for the ITS, trnL-F, and trnH-psbA sequences downloaded from GenBank and used in the phylogenetic analyses. New sequences generated for this study are shown in APPENDIX 1.

#### Supplementary_TableS1

TableS1. Species ocurrence data of Stenodraba used in the niche analyses

#### Supplementary_TableS2

TABLE S2. Features of the DNA regions included in the phylogenetic analyses within Eudemeae.

#### Supplementary_TableS3

TABLE S3. Loadings on the first two components obtained from the principal component analysis (PCA-env) using bioclimatic variables extracted from the study area (50 km-buffered minimum convex polygon enclosing the Stenodraba species distribution).

#### Supplementary_TableS4

TABLE S4. Number of ocurrences (n), area under the receiver operating characteristic curve (AUC), variable contribution (%), and permutation importance (%) obtained in the MAXENT modeling for each Stenodraba clade. Variable contributions to overall performance greater than 25% are indicated in bold. BIO3: Isothermality, BIO4: Temperature Seasonality, BIO8: Mean Temperature of Wettest Quarter, BIO12: Annual Precipitation, BIO15: Precipitation Seasonality.

#### Supplementary_FIGS1

FIG. S1. Bayesian 50% majority-rule consensus tree from 3,002 trees generated by Bayesian inference with MrBayes for Brassicaceae. A. ITS dataset. B. trnL-F dataset. Tribes are indicated to the right of the boxes. Thick branches indicate >0.90 Bayesian posterior probability. The white arrows indicate phylogenetic position of Stenodraba, while black arrows show the position of Pennellia (tribe Halimolobeae) and Weberbauera (tribe Thelypodieae) .

#### Supplementary_FIGS2

FIG. S2. Phylogenetic placement of Stenodraba within tribe Eudemeae, and including indel data in the analyses (nucleotide substitution + indel data). A ̶ C. Bayesian 50% majority-rule consensus trees from 6,002 trees generated by Bayesian inference with MrBayes. A. nrITS dataset. B. cpDNA dataset (trnL-F/trnH-psbA). C. Concatenated cpDNA + nrITS dataset (substitution + indel data). Values on branches correspond to Bayesian posterior probability (%). The start indicates the crown node of tribe Eudemeae.

#### Supplementary_FIGS3

FIG. S3. Phylogenetic placement of Stenodraba within tribe Eudemeae. A. Maximum clade credibility species tree (MCCT) estimated from nuclear ribosomal ITS and two chloroplast DNA regions (trnL-F, trnH-psbA) using the multispecies coalescent method implemented in *BEAST. Bayesian posterior probability values >50 % are given at each node. B. Filtered supernetwork generated in SplitsTree using 1000 Bayesian posterior trees per each nuclear ITS and plastid (trnL-F/trnH-psbA) dataset, and filtering the splits to show only those present in a minimum of 35% input trees. The start indicate the crown node of tribe Eudemeae.

#### Supplementary_FIGS4

FIG. S4. Results from the niche overlap analyses in the space environmental (E) space obtained using the PCA-env method. A. Correlation circle showing the contribution of environmental variables on the first two axes of the PCA-env, and the percentage of inertia explained by the two axes. B ̶ C. Values for the first two principal components obtained in the PCA-env analyses in the studied area. B. Values for the PC1. C. Values for the PC2. Red dots represent specimen records for species included in the clade A (S. imbricatifolia, S. lagunae, S. suffruticosa), while blue dots represent occurrences for species of clade B (S. lechleri, S. parvifolia, S. chillanensis, S. colchaguensis, S. stenophylla).

#### Supplementary_FIGS5

FIG. S5. Climatic niches of Stenodraba species in the environmental space produced by the two main axes of the PCA-env. For each species, the gray to black shading represents the grid cell density of the clade occurrence (black being the highest density). The first dashed line represents the 50% of the available environment and the solid line represents the 100 %. A. S. lechleri. B. S. parvifolia. C. S. chillanensis. D. S. colchaguensis. E. S. imbricatifolia. F. S. lagunae. G. S. stenophylla. H. S. suffruticosa.

#### Supplementary_FIGS6

FIG. S6. Values of variables that most contribute in the MaxEnt models for clades A and B of Stenodraba along the distribution area. A. BIO3 (Isothermality). B. BIO4 (Temperature Seasonality). C. BIO8 (Mean Temperature of Wettest Quarter). D. BIO12 (Annual Precipitation). E. BIO15 (Precipitation Seasonality). Red and blue dots represent specimen occurrences of clades A and B, respectively, used for the SDM analyses. F. Response curves for variables used in the MaxEnt models of clades A and B. Response curves show the ranges in environmental conditions that are more favorable for the distribution of the species of each clade. The x-axis of the variables represents their ranges for the complete study area, while the y-axis represents the predicted suitability of the focus variable when all of the other variables are set to their average. Red line represents the mean response while blue shaded area represents ± 1 standard deviation.

#### Supplementary_FIGS7

FIG. S7. Stenodraba species. A. Stenodraba chillanensis (Zuloaga 15235, SI), habit. B. Stenodraba colchaguensis (Zuloaga 12539, SI), habit. C-D. Stenodraba parvifolia (Zuloaga 15583, SI). C. Habit. D. flowers. Photographs by: A, C, D, D. Salariato; B, C. Zannotii.

#### ITS alignment BRASSICACEAE

ITS alignment of Brassicaceae dataset for family-level analyses.

ITS_BRASSICACEAE.fas

#### trnL-F alignment BRASSICACEAE

trnL-F alignment of Brassicaceae dataset for family-level analyses.

trnL-F_BRASSICACEAE.fas

#### Maximum likelihood tree ITS BRASSICACEAE

Maximum likelihood tree for Brassicaceae generated with RAxML using the ITS dataset.

BRASSICACEAE_ML_ITS.tre

#### Maximum likelihood tree trnL-F BRASSICACEAE

Maximum likelihood tree for Brassicaceae generated with RAxML using the trnL-F dataset

BRASSICACEAE_ML_trnL-F.tre

#### Bayesian 50% majority-rule consensus tree ITS BRASS

Bayesian 50% majority-rule consensus tree from 3002 trees generated by Bayesian inference with MrBayes and using the ITS dataset for Brassicaceae

BRASSICACEAE_BI_ITS.tre

#### Bayesian 50% majority-rule consensus trees trnL-F BRASS

Bayesian 50% majority-rule consensus trees from 3002 trees generated by Bayesian inference with MrBayes and using the trnL-F dataset for Brassicaceae

BRASSICACEAE_BI_trnL-F.tre

#### ITS_EUDEMEAE

ITS alignment of tribe Eudemeae dataset including Stenodraba accessions.

#### trnL-F_EUDEMEAE

trnL-F alignment of tribe Eudemeae dataset including Stenodraba accessions.

#### trnH-psbA_EUDEMEAE

trnH-psbA alignment of tribe Eudemeae dataset including Stenodraba accessions.

#### ITS_EUDEMEAE_Nexus

Nexus file with the ITS matrix of tribe Eudemeae including Stenodraba accessions.

ITS_EUDEMEAE.nex

#### trnL-F_EUDEMEAE_Nexus

Nexus file with the trnL-F matrix of tribe Eudemeae including Stenodraba accessions.

trnL-F_EUDEMEAE.nex

#### trnH-psbA_EUDEMEAE_Nexus

Nexus file with the trnH-psbA matrix of tribe Eudemeae including Stenodraba accessions.

trnH-psbA_EUDEMEAE.nex

#### RAxML_bestTree.ITS

Maximum likelihood tree for analyses within tribe Eudemeae generated with RAxML and using the ITS dataset.

#### RAxML_bestTree.trnL-F

Maximum likelihood tree for analyses within tribe Eudemeae generated with RAxML and using the trnL-F dataset.

#### RAxML_bestTree.trnH-psbA

Maximum likelihood tree for analyses within tribe Eudemeae generated with RAxML and using the trnH-psbA dataset

#### RAxML_bestTree.cpDNA

Maximum likelihood tree for analyses within tribe Eudemeae generated with RAxML and using the concatenated trnL-F + trnH-psbA datasets

#### RAxML_bestTree.ITS+cpDNA

Maximum likelihood tree for analyses within tribe Eudemeae generated with RAxML and using the concatenated ITS + trnL-F + trnH-psbA datasets

RAxML_bestTree.ITS_cpDNA.tre

#### MrBayes50_ITS

Bayesian 50% majority-rule consensus trees from 6002 trees generated by Bayesian inference with MrBayes for tribe Eudemeae and the nrITS dataset.

BI_ITS.tre

#### MrBayes50_trnL-F

Bayesian 50% majority-rule consensus trees from 6002 trees generated by Bayesian inference with MrBayes for tribe Eudemeae and the trnL-F dataset.

BI_trnL-F.tre

#### MrBayes50_trnH-psbA

Bayesian 50% majority-rule consensus trees from 6002 trees generated by Bayesian inference with MrBayes for tribe Eudemeae and the trnH-psbA dataset.

BI_trnH-psbA.tre

#### MrBayes50_cpDNA

Bayesian 50% majority-rule consensus trees from 6002 trees generated by Bayesian inference with MrBayes for tribe Eudemeae and the concatenated trnLF + trnH-psbA datasets.

BI_cpDNA.tre

#### MrBayes_ITS+cpDNA

Bayesian 50% majority-rule consensus trees from 6002 trees generated by Bayesian inference with MrBayes for tribe Eudemeae and the concatenated ITS+ trnLF + trnH-psbA datasets.

BI_ITS+cpDNA.tre

#### ITS+indel_MrBayes

Alignment matrix of ITS sequences for Bayesian inference with MrBayes including indel data

#### cpDNA+indel_MrBayes

Alignment matrix of concatenated trnL-F + trnH-psbA datasets for Bayesian inference with MrBayes including indel data

#### ITS+cpDNA+indel_MrBayes

Alignment matrix of concatenated ITS + trnL-F + trnH-psbA datasets for Bayesian inference with MrBayes including indel data

#### MrBayes50tree_ITS+indel

Bayesian 50% majority-rule consensus trees from 6002 trees generated by Bayesian inference with MrBayes for tribe Eudemeae and the nrITS dataset + Indel data.

BI_ITS+indel.tre

#### MrBayes50tree_cpDNA+indel

Bayesian 50% majority-rule consensus trees from 6002 trees generated by Bayesian inference with MrBayes for tribe Eudemeae and the concatenated trnLF + trnH-psbA datasets + Indel data.

BI_cpDNA+indel.tre

#### MrBayes50tree_ITS+cpDNA+indel

Bayesian 50% majority-rule consensus trees from 6002 trees generated by Bayesian inference with MrBayes for tribe Eudemeae and the concatenated ITS+ trnLF + trnH-psbA datasets + Indel data.

BI_ITS+cpDNA+indel.tre

#### PrimaryConcordanceTree ITSvscpDNA Bucky

Primary concordance tree from the BUCKy analyses using the 6,002 trees from the ITS and the cpDNA datasets obtained in the MrBayes analyses. The discordance parameter (α) was set to 1, 10, and 100, however results from concordance analyses varying the discordance prior (α) had no effect on topology or concordance.

PCT_Bucky.tre

#### xml BEAST for species tree analysis

xml file for species tree analysis under the multispecies coalescent model in BEAST 2.4.4

STENO_species.xml

#### MCC Species Tree BEAST

Maximum clade credibility species tree (MCCT) estimated from nuclear ribosomal ITS and two chloroplast DNA regions (trnL-F, trnH-psbA) using the multispecies coalescent method implemented in *BEAST.

MCCSpeciesTree.tre