How to perform burden analysis with PLINK?

[Caterina]

I saw this for plink1.7 but not sure if it exists for 1.9 or 2.

https://zzz.bwh.harvard.edu/plink/data.shtml#sets

Basically I want to obtain number of mutations (SNP=alt) within a gene for each person.

How could I do that?

The output could be either 

indID number_of_mut

ind1 1

ind2 2

ind3 0

ind4 1

Or the sum of them

individuals with 1 mutation in geneA = 2

individuals with 2 mutations in geneA = 1

individuals with 0 mutations in geneA = 1

Thank you!

--freqx provides this at the variant level I would like it at the "gene" level. 

[Master thesis]

I forgot to mention that I also need the phasing info, and whether it is homozygous or heterozygous. 

Thank you!

--
You received this message because you are subscribed to a topic in the Google Groups "plink2-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plink2-users/VQqlk3fJWLE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to plink2-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/plink2-users/d22b073e-6993-4e09-ba57-d2700d918455n%40googlegroups.com.

[Christopher Chang]

1. You need to know which allele to count for each SNP.  If you do, you can provide this information to plink 2.0 --score, in a file that looks something like

snp1  A  1

snp2  C  1

snp3  T  1

snp4  A  1

...

Then "--score <name of the file described above> no-mean-imputation cols=+scoresums" should provide the totals you want in the last column of the output file.

2. If you have phased data and want a separate count for each side, there is currently no plink command to compute that directly.  (One workaround is to create a file with twice as many samples, where each new sample contains homozygous genotypes for one of the original sample-sides.  I'm not aware of a preexisting program that does this, but it's pretty straightforward to write a script that performs this transformation to a VCF file.)

[Caterina]

Thank you! Is it possible to do this for two genes in combination?

indID mut_in_gene_A mut_in_gene_B

ind1 1 2

ind2 2 3

ind3 0 1

ind4 1 0

[Master thesis]

Or should I just run --score twice for each gene?

To view this discussion on the web visit https://groups.google.com/d/msgid/plink2-users/d198001b-8004-4aae-b67b-a59c462972fcn%40googlegroups.com.

[Caterina]

Could you please explain me the difference between dosagesum and scoresums?

Wouldn't they be the same in my case?

[Christopher Chang]

1. --score-list provides one way to compute scores for multiple genes at once.

2. With --score on a single gene, yes, dosagesum and scoresums would have the same values because every multiplier is 1.  (If any third-column value was not 1, they would no longer be guaranteed to have the same value.)  But with --score-list on multiple genes, there is no dosagesum column.