Further information on the linkdatagen software package can be found here: http://bioinf.wehi.edu.au/software/linkdatagen/
Further information on the linkdatagen software package can be found here: http://csg.sph.umich.edu/abecasis/merlin/ You will need to compile locally. If you are on Nimbus, this is easy because you have admin privs, and it will automatically be installed into your PATH. If you do not have admin priovs, you will need to choose another install directory and then add it to your PATH if it's not there already.
We use linkdatagen to manipulate genetic variants from either NGS/MPS data, or SNP-arrays. For NGS data, linkdatagen extracts informative variants from the sequencing data to use as datapoints for linkage analysis with Merlin.
cd into your desired directory (i.e. /data/ on nimbus)
git clone https://github.com/SarahBeecroft/Linkage.git
cd Linkage
chmod 777 *.sh *.pl
./DOWNLOADS.sh
then install Merlin (instructions provided at end of DOWNLOADS.sh
You will need to create/upload the input files for the linkage program to use. The sequencing data input will be the bam files provided by the sequencing facility. These need to be uploaded into the Linkage working directory via filezilla or WinSCP.
Other input files are simply text files- the familyID.ped, familyID.ws and familyID.vcflist.txt files, where 'familyID' is a prefix of your choosing. Spaces, slashes and other special characters are not allowed in the prefix! Make sure to name the files exactly as outlined here (case sensitive), otherwise the program will not work.
These files are crucial to get right! Any small errror may confound your results without it being obvious. It's vital to keep the sample names within these files in a consistent order throughout. E.g. Dad, Mum, Bro, Sis, Bro2 in order throughout all the files.
This is a representation of the pedigree. The values must be separated by tabs or spaces.
Although pedigrees can become quite complex, all the information that is necessary to reconstruct individual relationships in a pedigree file can be summarized in five items: a family identifier, an individual identifier, a link to each parent (if available) and finally an indicator of each individual's sex. We then add whether an individual has the disease or not (affection status).
As an example of how family relationships are described, we will construct a pedigree file for a small pedigree with two siblings, their parents and maternal grand-parents.
For this simple pedigree, the five key items take the following values:
FAMILY PERSON FATHER MOTHER SEX AFFECTION_STATUS
example granpa unknown unknown m unaffected
example granny unknown unknown f affected
example father unknown unknown m unaffected
example mother granpa granny f affected
example sister father mother f affected
example brother father mother m affected
These key values constitute the columns of the pedigree file. Because of restrictions in early genetic programs, text identifiers are usually replaced by unique numeric values.
Male = 1, Female = 2, Unknown = 0. Unaffected = 1, Affected = 2, Unknown = 0.
In this pedigree, granpa and granny are what we call 'founders'- they are the first generation in the pedigree and therefore are allowed to have 'unknown' parents (i.e. mum and dad values of 0). father is also a founder, because he married into the family. All 'non-founders' require their parents to be specified by using the sample number of the mum or dad (in this case, mother has granpa and granny specified as her parents- 1 and 2 respectively
After replacing each identifier with unique identifier and recoding sexes/affection status, this is what a basic space-delimited pedigree file would look like:
1 D18-0001 0 0 1 1
1 D18-0002 0 0 2 2
1 D18-0003 0 0 1 1
1 D18-0004 D18-0001 D18-0002 2 2
1 D18-0005 D18-0003 D18-0004 2 2
1 D18-0006 D18-0003 D18-0004 1 2
A pedigree file can include multiple families. Each family can have a unique structure, independent of other families in the dataset.
*It's important to note that we need
However, we will not always have sequencing data for everyone in the family! We need the .ws file to tell linkdatagen which samples we have data on, and which we don't.
The .ws file is a single line of N values, where N equals the number of individuals in the pedfile. The values must be separated by tabs or spaces. The kth value indicates where to find genotypes for the individual listed in the kth row of the pedigree file. The value 0 indicates that the corresponding individual is not genotyped.
Using the above example, if we had sequencing data for granny, mother and sister only, the ws would be:
0 1 0 2 3 0
The vcflist.txt file is the full path of the .HM.vcf files that will be generated by linkdatagen. Basically, the format is
/path/to/Linkage/working/dir/sampleID1.HM.vcf
/path/to/Linkage/working/dir/sampleID2.HM.vcf
/path/to/Linkage/working/dir/sampleID3.HM.vcf
Make sure the listed files are in the same order as the ped file and the ws file (obviously skipping any sampleIDs that didn't get sequenced).
Usage is ./samtools_gen.HM.vcf This will make the HM.vcf files from your bam files. These only need to be generated once per bam, but takes the longest amount of time. You will need to provide a space seperated list of the sample IDs for the bam files. Note that the program is assuming your bams are named sample_id.bam so you may need to rename the bams to fit this convention. Of course, always leave the original bam files living on IRDS etc unchanged Only edit your copies.
This program will do the rest of the linkage work for you once you've done the samntools step. It will ask you for the input files outlined above. The program will then start running. This may take some time. When it is finished, you will have two new directories. The one of interest is familyID.HapMap2_merlin, which has a sub-directory called genome. Within genome, the files you want are the parametric.pdf and parametric.tbl.