Skip to content

Latest commit

 

History

History
197 lines (145 loc) · 10.7 KB

usage.md

File metadata and controls

197 lines (145 loc) · 10.7 KB

Usage

This section describes how to use Structure_threader.

Structure_threader can be executed via two main modes.

  • run: The main execution mode that performs the parallel execution of the external structuring program, calculates the best K values (ALStructure currently does not support finding the best K) and generates the plot files
  • plot: This execution mode will only generate new plot files from the output files of the structuring program.

run mode

Using the run mode, the program currently takes the following arguments:

  • I/O arguments:
    • Input file (-i)
    • Output directory (-o)
    • Path to parameters_file (path to mainparams [will also assume extraparams] or parameters.txt; --params)
  • Individual/Population identification options:
    • Path to popfile (--pop) [See below for more information]
    • Path to indfile (--ind) [See below for more information]
  • External program location - you have to pass one and only one of the following arguments:
    • STRUCTURE location (if you want to run STRUCTURE; -st)
    • fastStructure location (if you want to run fastStructure; -fs)
    • MavericK location (if you want to run MavericK; -mv)
    • ALStructure_wrapper.R location (if you want to run ALStructure; -als)
  • Number of K - you have to pass one and only one of the following arguments:
    • K (To test all values of "K" from 1 to "K"; -K)
    • Klist (To test all values of "K" in the provided list; -Klist)
  • Replicates (ignored for fastStructure, MavericK and ALStructure; -R)
  • Number of threads to use (-t)
  • Q-matrix plotting options:
    • Disable plot drawing (--no_plots)
    • Force plotting the given values together (--override_bestk)
    • Draw the plots only in grayscale (-bw)
  • Other options
    • Enable logging - useful when problems arise (--log)
    • Do not run the BestK tests (--no-tests)
    • Add extra arguments to pass to the wrapped program (--extra_opts) [Example: prior=logistic seed=123]
    • Define a random seed starting value (--seed) [default:1235813]

Example run:

structure_threader run -K Ks -R replicates -i infile -o outpath -t num_of_threads -st path_to_structure

Where -K is the number of "Ks" to run, -R is the number of replicate runs for each value of "K", -i is the input file for STRUCTURE, -o is the directory where the output results should be stored, -t is the number of threads to use and -p the path for the STRUCTURE binary.

The program should be run in the same directory where the files "mainparams" and "extraparams" for your STRUCTURE run are placed. Please see Installation for information on how to achieve this. Alternatively, you can specify the path to where your parameter files (mainparams and extraparams or parameters.txt) and Structure_threader will read parameters from the specified location. This can be achieved using the --params switch.

plot mode

Using the plot mode, the program currently takes the following arguments:

  • Main plotting options:

    • Directory where the Qmatrix files you want to plot are located. In the case of MavericK point to where the directories named mav_KX are located (-i)

    • External program or format of the output files. This can be 'structure', 'fastStructure' or 'maverick'. (-f)

    • The K values that you want to plot. Each individual K value that is provided will be plotted individually and in the end, a comparative plot will all K values will also be generated. (-K; Example: -K 2 3 4)

    • Directory where the plots will be saved to. By default they will be generated in the current working directory (-o)

  • Individual/Population identification options:

    • Path to popfile (--pop) [See below for more information]
    • Path to indfile (--ind) [See below for more information]

  • Extra plotting options:

    • Do not use colors when drawing the plots (-bw)
    • Use individual sample labels even when population labels are available (--use-ind-labels)

Example run:

structure_threader plot -i fS_run -f fastStructure -K 2 3 4 -o 2_4_plots --ind indfile.txt

Here, Structure_threader will search the current directory for all FastStructure output files that start with the "fS_run" string, as specified by the "-i" option. The "-f" option specified the fastStructure format of the output files. The "-K" option specified which K values should be plotted (Note: If any of the provided K values do not exist, they are ignored). Using the "-o" option the plots will be generated into the "2_4_plots" directory. Finally, we also provide an indfile using the "--ind" option.

params mode

Using the params mode, Structure_threader generate a skeleton mainparams and extraparams that you should edit to facilitate STRUCTURE runs. Most options have been preset to a commonly used default value, but some of them are set to "CHANGEME" since providing a default value here makes no sense, since it depends on each dataset. The params mode takes only one option:

  • Output directory (path to where the skeleton parameter files should be written; -o)

Using a popfile

Structure_threader can build your structure plots with labels and in a specified order. For that you have to provide a "popfile" (--pop option). This file consists of the following 3 columns: "Population name", "Number of individuals in the population", "Order of the population in the plot file". Here is an example:

Location_1  20  1
Location_2  21  2
Location_3  11  3

This example file contains 3 populations, with 20, 21 and 11 individuals. This means that the first 20 individuals are from "Location_1", the next 21 individuals are from "Location_2" and so on. The numbers "1", "2" and "3" will be the order of the populations in the plot file. If you want to draw the plot in a different order than what was provided on the input file, you have to reorder the lines. For the sake of the example, let's say that you wish to plot your data, switching the place of Locations 3 and 2. The input file would look like this:

Location_1  20  1
Location_3  11  3
Location_2  21  2

You can use any order you like using this scheme. Also note that the "split bars" that split the populations in your plot will correspond to the number provided in column 2.

Using an indfile

The indfile works in a similar fashion to the popfile but discriminates each individual sample in the output file. This file can be provided with the --ind option and is particularly useful when you have no knowledge of the populations or when individuals from the same population are not clustered together in the input file.

This file can have between one and up to three columns:

  1. Individual sample name
  2. Population name
  3. Order of the population in the plot

Only the first column is mandatory and if you provide an indfile with a single column, the resulting plot will contain the individual samples as the x-axis labels. Such indfile could be simply:

Ind1
Ind2
Ind3

Note that this will not re-order the columns in your plot - this will simply change the individual labels! In order to avoid problems, triple check that the indfile order is the same as in your original input file.

If you provide the second column with the population name, the x-axis labels of the resulting plots will only display the population names. However, the .html plots will also display the name of the individual sample names when hovering the mouse over their respective bars. An example of a two column indfile:

Ind1    PopA
Ind2    PopB
Ind3    PopA

Like previously, this option will also not reorder the plot columns! It simply changes the labels.

Adding a third column will allow you to change the order of the populations in the generated plots. In this example:

Ind1    PopA    2
Ind2    PopB    1
Ind3    PopA    2

The individuals of "PopB" will appear first, and then the individuals of "PopA".

Note: While popfiles and indfiles are a simple and convenient way to change the population order, they are not meant to change individual order (albeit individuals will be moved in the plot when population order changes) in the plot. If you are not happy with the individual order in your plot, the best way to change it is in the input file. I can implement a "plot order changer" routine if there is enough demand for it, though (you can provide feedback by commenting on the issue either on gitlab or github).

fastStructure Warning:

Keep in mind that fastStructure can take input in two distinct file formats: Plink and structure. In order to use the PLINK format, three files are required:

  • file.bed
  • file.fam
  • file.bim

You can enter any of them (but just one of them) as the input file and Structure_threader will assume the other two exist in the same path. If the input file specified by the -i switch in Structure_threader has an extension different from either of the three mentioned above, Structure_threader will assume th input is in the STRUCTURE format, which has some peculiarities: fastStructure requires your input file to have each individual represented in two rows (one for each allele), and six "bogus" columns before the actual data. No Header is allowed. Here is a short example:

Ind1    col1  col2  col3  col4  col5 1    3   1   4
Ind1    col1  col2  col3  col4  col5 1    3   2   4
Ind2    col1  col2  col3  col4  col5 1    2   1   4
Ind2    col1  col2  col3  col4  col5 1    2   1   3

ALStructure Warning:

Keep in mind that ALStructure can take input in two distinct file formats: Plink and VCF. In order to use the PLINK format, three files are required:

  • file.bed
  • file.fam
  • file.bim

You can enter any of them (but just one of them) as the input file and Structure_threader will assume the other two exist in the same path.

In order to use a VCF formatted file, it is only required that you point at it with the -i switch. Note that when using this format, Structure_threader will create a new file, with the same name and in the same PATH as your VCF file, but with a .tsv extension, which is ready to be parsed by ALStructure.

Don't forget to look at the Output section for information on how the data is presented after a successful (or not) run.

Using MavericK:

MavericK is exhaustively documented. You can find the full manual here, along with other useful material to make the most of the software.