/tools/rgenetics/rgGRR.xml
https://bitbucket.org/cistrome/cistrome-harvard/ · XML · 95 lines · 76 code · 19 blank · 0 comment · 0 complexity · 5f96b2dbcad3bfcf4c5ada6955226f35 MD5 · raw file
- <tool id="rgGRR1" name="GRR:">
- <description>Pairwise Allele Sharing</description>
- <command interpreter="python">
- rgGRR.py $i.extra_files_path/$i.metadata.base_name "$i.metadata.base_name"
- '$out_file1' '$out_file1.files_path' "$title" '$n' '$Z'
- </command>
- <inputs>
- <param name="i" type="data" label="Genotype data file from your current history"
- format="ldindep" />
- <param name='title' type='text' size="80" value='rgGRR' label="Title for this job"/>
- <param name="n" type="integer" label="N snps to use (0=all)" value="5000" />
- <param name="Z" type="float" label="Z score cutoff for outliers (eg 2)" value="6"
- help="2 works but for very large numbers of pairs, you might want to see less than 5%" />
- </inputs>
- <outputs>
- <data format="html" name="out_file1" label="${title}_rgGRR.html"/>
- </outputs>
- <tests>
- <test>
- <param name='i' value='tinywga' ftype='ldindep' >
- <metadata name='base_name' value='tinywga' />
- <composite_data value='tinywga.bim' />
- <composite_data value='tinywga.bed' />
- <composite_data value='tinywga.fam' />
- <edit_attributes type='name' value='tinywga' />
- </param>
- <param name='title' value='rgGRRtest1' />
- <param name='n' value='100' />
- <param name='Z' value='6' />
- <param name='force' value='true' />
- <output name='out_file1' file='rgtestouts/rgGRR/rgGRRtest1.html' ftype='html' compare="diff" lines_diff='350'>
- <extra_files type="file" name='Log_rgGRRtest1.txt' value="rgtestouts/rgGRR/Log_rgGRRtest1.txt" compare="diff" lines_diff="170"/>
- <extra_files type="file" name='rgGRRtest1.svg' value="rgtestouts/rgGRR/rgGRRtest1.svg" compare="diff" lines_diff="1000" />
- <extra_files type="file" name='rgGRRtest1_table.xls' value="rgtestouts/rgGRR/rgGRRtest1_table.xls" compare="diff" lines_diff="100" />
- </output>
- </test>
- </tests>
- <help>
- .. class:: infomark
- **Explanation**
- This tool will calculate allele sharing among all subjects, one pair at a time. It outputs measures of average alleles
- shared and measures of variability for each pair of subjects and creates an interactive image where each pair is
- plotted in this mean/variance space. It is based on the GRR windows application available at
- http://www.sph.umich.edu/csg/abecasis/GRR/
- The plot is interactive - you can unselect one of the relationships in the legend to remove all those points
- from the plot for example. Details of outlier pairs will pop up when the pointer is over them. e found by moving your pointer
- over them. This relies on a working browser SVG plugin - try getting one installed for your browser if the interactivity is
- broken.
- -----
- **Syntax**
- - **Genotype file** is the input pedigree data chosen from available library Plink binary files
- - **Title** will be used to name the outputs so make it mnemonic and useful
- - **N** is left 0 to use all snps - otherwise you get a random sample - much quicker with little loss of precision > 5000 SNPS
- **Summary**
- Warning - this tool works pairwise so slows down exponentially with sample size. An LD-reduced dataset is
- strongly recommended as it will give good resolution with relatively few SNPs. Do not use all million snps from a whole
- genome chip - it's overkill - 5k is good, 10k is almost indistinguishable from 100k.
- SNP are sampled randomly from the autosomes - otherwise parent/child pairs will be separated by gender.
- This tool will estimate mean pairwise allele shareing among all subjects. Based on the work of Abecasis, it has
- been rewritten so it can run with much larger data sets, produces cross platform svg and runs
- on a Galaxy server, instead of being MS windows only. Written in is Python, it uses numpy, and the innermost loop
- is inline C so it can calculate about 50M SNPpairs/sec on a typical opteron server.
- Setting N to some (fraction) of available markers will speed up calculation - the difference is most painful for
- large subject N. The real cost is that every subject must be compared to every other one over all genotypes -
- this is an exponential problem on subjects.
- If you don't see the genotype data set you want here, it can be imported using one of the methods available from
- the Rgenetics Get Data tool.
- -----
- **Attribution**
- Based on an idea from G. Abecasis implemented as GRR (windows only) at http://www.sph.umich.edu/csg/abecasis/GRR/
- Ross Lazarus wrote the original pdf writer Galaxy tool version.
- John Ziniti added the C and created the slick svg representation.
- Copyright Ross Lazarus 2007
- Licensed under the terms of the LGPL as documented http://www.gnu.org/licenses/lgpl.html
- </help>
- </tool>