SDMs with tidymodels | Accessing the data for this vignette: how to use rgbif | Preparing your data | Map projection | Thinning step | Fit the model by cross-validation | Projecting to the present | Projecting to the future | Dealing with extrapolation | Visualising the contribution of individual variables | Repeated ensembles

Graphs and f-statistics | Graphs and f2 | Graphs and f4 | f4 ratio test | Graph as a function | Identifiability | How it works | f-statistics vs SFS | Modeling demographic history | Distinguishing between a large number of alternative models requires a large amount of data | Incorporating prior information | f-statistics, qpAdm, and admixture graphs | Graphs in ADMIXTOOLS 2 | Valid graphs | Creating and loading admixture graphs | Modifying admixture graphs | Fitting a single graph | Initial weights | Regularization terms | Edge constraints | Drift edges fitted at zero | Confidence intervals | Why f3? | Exploring different graphs | Fully automated graph exploration | Optimizing worst residual instead of score | Constraining the search space | Semi-automated graph exploration | Comparing the fits of different graphs | Out-of-sample scores | Bootstrap-resampled graph fits | Choosing the number of admixture events | Summarizing graphs | Number of admixing sources | Order of events | Non-zero f4-statistics | Simulating under an admixture graph | msprime

Why LEGOFIT | Write: a graph to a .lgo file | Fit: run LEGOFIT | Read: the fitted edge tibble | Reading a real .lgo file | Availability

Introduction | f-statistics basics | f~2~ in ADMIXTOOLS 2 | f~3~ and qp3Pop | f~4~ and qpDstat | F~ST~ | qpWave and qpAdm | Running many models | Pairwise cladality tests | Rotating outgroups | Many qpadm models | Sweeps over target x source x right | qpGraph | More about graphs | Comparing results between ADMIXTOOLS and ADMIXTOOLS 2

Genotype data formats | PLINK | PLINK 2 (PFILE) | EIGENSTRAT/PACKEDANCESTRYMAP | Reading genotype files | Extracting populations | Combining data sets | Computing allele frequencies | Converting PACKEDANCESTRYMAP to PLINK | f2 cache metadata | Admixture graph formats | Edge list format | igraph format | Original ADMIXTOOLS format | DOT format

f~2~ | f~4~ | Biases | Bias due to inaccurate allele frequency estimates | Bias due to missing data | Bias due to SNP ascertainment | Common vs rare SNPs | f-statistics in ADMIXTOOLS 2 | Allele frequency products | Extracting f~2~ for a large number of populations | Extracting f~2~ for arbitrary populations

Plot admixture graph | Admixture graph on a map | All samples on a map

qpWave | qpWave and f4 | Estimating the number of independent gene flows | Estimating the rank of X | qpAdm | Modelling an admixed population | qpadm in practice | Diagnosing collinear right populations | Multiple models: qpadm_multi and qpadm_sweep | qpwave-style runs against genotype data

Quantifying uncertainty | Standard errors and hypothesis testing | Jackknife and bootstrap | Blocked jackknife and bootstrap | Resampling SNPs | Testing whether two graphs fit significantly differently | Resampling individuals

What's parallelized | Choosing a future plan | Controlling how many cores are used | Parallelization on a compute cluster | When parallelization doesn't help

Quality control for SNP datasets | Read data into gen_tibble format | Quality control for individuals | Quality control for loci | Linkage Disequilibrium | Save | Grouping data | Grouped functions

Using multi-level factors as predictors | Exploring models with DALEX | Different recipes for certain models | The initial split | Stack ensembles

Tidy data in population genetics | The gen_tibble | Standard dplyr verbs to manipulate the tibble | Using verbs on loci | Grouping individuals in populations | Functions applying to all pairwise and nwise combinations of individuals or populations | Reading data | Merging data | Imputation | Saving data and updating backingfiles | Ordering loci

Creating a gen_tibble | Quality control | Impute | PCA | Plot with ggplot2

An example workflow with real data | Map | PCA | Explore population structure with DAPC | Clustering with ADMIXTURE | Handling Q matrices

Installation | An overview of CRS codes | Spatial objects in R | Choosing a CRS

File management and reading data: | Quality control: | Handling linkage | Quality control for relatedness (KING) | Merging datasets: | Analysis:

NAs in the data | Recipes and the response variable | Using the desired formula with GAM | When only some splits fail

SDMs with tidymodels for palaeo data | Preparing your data | Fit the model by crossvalidation | Projecting to other times

Overview of datasets available in pastclim

Downscaling | Delta downscaling a dataset in pastclim | An example for one variable | Computing the bioclim variables

Install the library | Download the data | Get climate for locations | Get climate for a region | Cropping | Working with biomes and ice sheets | Adding locations to region plots | Set the samples within the background | Random sampling of background | Downscaling

Present reconstructions | WorldClim | CHELSA | Future projections

Formatting a custom dataset for pastclim | An example: the Trace21k-CHELSEA | Making the data available to others

geoGraph: walking through the geographic space using graphs. | First steps | Installing the package | Data representation | gGraph objects | gData objects | Using geoGraph | Importing geographic data | Visualizing data | Plotting gGraph objects | Zooming in and out, sliding, etc. | Plotting gData objects | Editing gGraphs | Changing the global connectivity of a gGraph | Changing local properties of a gGraph | Extracting information from GIS shapefiles | Finding least-cost paths

Thoughts on Maier, Flegontov et al. | How to prevent overfitting | More thoughts on admixture graphs | What can be concluded if multiple different admixture graph models fit around equally well? | If we don't have enough data to fit complex models, can we just fit simpler models? | When are admixture graph models adequate? | What other options are there? | Correlation, causality, and admixture graphs | More reasons tread carefully | Researcher degrees of freedom | SNP ascertainment | Conclusion