qtl2cl

Karl Broman

qtl2cl is part of R/qtl2. It provides a command-line interface to a restricted set of R/qtl2 functions.

---

Installation

Install R/qtl2cl from GitHub using the remotes package.

install.packages("remotes")

Then install R/qtl2cl with install_github:

remotes::install_github("rqtl/qtl2cl")

The package dependencies (including R/qtl2, R/qtl2convert, and optparse) will also be installed.

The command-line script will be located at $R_LIBS/qtl2cl/scripts/qtl2cl where $R_LIBS is the path to the R packages within your R installation. You may want to add this to your PATH. Once you've installed the R/qtl2cl package, you can use the following to find the path to the qtl2cl script.

Rscript -e "system.file('scripts', 'qtl2cl', package='qtl2cl')"

---

Usage

Currently, the command-line interface to R/qtl2 can do the following things.

Import cross data and save to RDS file

With the option --cross2rds, you can load data for a cross and save the resulting "cross2" object to an RDS file. For example, we can download the B6xBTBR intercross data and save it to a local RDS file.

qtl2cl --cross2rds --input=https://raw.githubusercontent.com/rqtl/qtl2data/master/B6BTBR/b6btbr.zip --output=b6btbr.rds

Calculate genotype probabilities

With the option --calc_genoprob, you can load data for a cross from an RDS file and calculate QTL genotype probabilities. Control of the calculations is through the following arguments:

• --step
• --off_end
• --stepwidth
• --error_prob
• --map_function
• --cores

It may also be important to write the genetic marker map, with pseudomarkers inserted, to an RDS file. This is accomplished by providing the --map_file argument.

Here's an example using the B6xBTBR intercross data. We've split the command across two lines so that it's easier to read.

qtl2cl --calc_genoprob --input=b6btbr.rds --output=b6btbr_probs.rds --map_file=b6btbr_map.rds \
   --step=0.5 --stepwidth=max --error_prob=0.002 --map_function=c-f

Convert genotype probabilities to allele dosages

With the option --genoprob_to_alleleprob, you can convert genotype probabilities to allele dosages. This is useful for performing a genome scan with an additive allele model. The allele dosages can also be used to calculate the kinship matrix.

Here's an example using the B6xBTBR intercross data.

qtl2cl --genoprob_to_alleleprob --input=b6btbr_probs.rds --output=b6btbr_aprobs.rds

Calculate kinship matrices

With the option --calc_kinship, you can calculate a kinship matrix from previously calculated genotype or allele probabilities. These are used in a linear mixed model to adjust for background polygenic effects. Control of the calculations is through the following arguments:

• --type (overall, loco, or chr)
• --use_grid_only
• --omit_x
• --use_allele_probs

Here's an example using the B6xBTBR intercross data.

qtl2cl --calc_kinship --input=b6btbr_aprobs.rds --output=b6btbr_kinship.rds

Grab X chromosome covariates

With the option --get_x_covar, you can grab the special X chromosome covariates needed for QTL analysis with some cross types. The input is a cross object saved as an RDS file. The result is saved to another RDS file.

Here's an example using the B6xBTBR intercross data.

qtl2cl --get_x_covar --input=b6btbr.rds --output=b6btbr_xcovar.rds

Perform single-QTL genome scan

With the option --scan1, you can perform a single-QTL genome scan. The minimal inputs are genotype (or allele) probabilities and phenotypes. You can also provide additive covariates, interactive covariates, special X chromosome covariates, kinship matrices, and weights. Each of these is provided as a file name, which may be for a CSV, RDS, JSON, or YAML file (except for the genotype probabilities, which must be in an RDS file). In each case, the contents should be a rectangle of numeric values with a header row and with the first column being a set of individual identifiers.

If an output file is provided, the results are saved as an RDS file. If no output file is provided, the results are printed to STDOUT as JSON. In the latter case, one would generally want to provide the genetic marker/pseudomarker map as an RDS file (via --map_file), created when running --calc_genoprob. Otherwise the LOD scores will be provided but without information about chromosomes and positions.

Here's an example using the B6xBTBR intercross data, with the phenotypes grabbed from the web and the X chromosome covariates derived above (with qtl2cl --get_x_covar). We've split the command across two lines, because the URL for phenotypes is so long.

In this case, the genome scan results will be saved to an RDS file, b6btbr_scan.rds.

qtl2cl --scan1 --genoprobs=b6btbr_aprobs.rds --Xcovar=b6btbr_xcovar.rds --output=b6btbr_scan.rds \
    --pheno=https://raw.githubusercontent.com/rqtl/qtl2data/master/B6BTBR/b6btbr_pheno.csv

Alternatively, if --output is not provided, the results are printed to STDOUT as a JSON object. In this case, it's best to provide --map_file (as used, to save the map object, when running --calc_genoprob), so that the JSON output includes chromosome and position information.

qtl2cl --scan1 --genoprobs=b6btbr_aprobs.rds --Xcovar=b6btbr_xcovar.rds --map_file=b6btbr_map.rds \
    --pheno=https://raw.githubusercontent.com/rqtl/qtl2data/master/B6BTBR/b6btbr_pheno.csv

---

License

Licensed under GPL-3.