search

andreyshabalin / MatrixEQTL

Matrix eQTL: Ultra fast eQTL analysis via large matrix operations
53 stars 16 forks source link

trans-QTL was defined as cis-QTL #12

Closed wdyyy closed 3 years ago

wdyyy commented 3 years ago

Dear Andrey Shabalin

I have met a strange problem while using MatrixEQTL. In my study, some results that should have been defined as trans QTL were defined as cis QTLs. In other words, some cis-QTLs have a distance over 1e6 bp between SNP and its gene. The distance is the minimum value of the absolute value of the distance between SNP and both ends of the gene.

    distance = min( abs( snp_position - gene_start ), abs( snp_position - gene_end )  )

there are some of the wrong QTLs:

qtl_type SNP_ID SNP_chr SNP_loc ref alt gene_ID gene_chr gene_start gene_end gene_strand beta t_stat p distance
cis rs7088262 chr10 49836108 A G BMS1P5 chr10 46737615 46762892 -;chr10 -0.23 -3.5 4.88E-04 3.073216
cis rs61838734 chr10 49681118 T C BMS1P5 chr10 46737615 46762892 -;chr10 0.14 3.74 1.91E-04 2.918226
cis rs72783091 chr10 49680879 A C BMS1P5 chr10 46737615 46762892 -;chr10 0.13 3.63 2.92E-04 2.917987
cis rs200554992 chr10 49680859 TA T BMS1P5 chr10 46737615 46762892 -;chr10 0.13 3.47 5.34E-04 2.917967
cis rs56342922 chr10 49680623 A G BMS1P5 chr10 46737615 46762892 -;chr10 0.14 3.68 2.46E-04 2.917731
cis rs7093726 chr10 49678147 T G BMS1P5 chr10 46737615 46762892 -;chr10 0.14 3.68 2.46E-04 2.915255
cis rs7069220 chr10 49678060 C T BMS1P5 chr10 46737615 46762892 -;chr10 0.13 3.51 4.59E-04 2.915168
cis rs3906889 chr10 49677182 G C BMS1P5 chr10 46737615 46762892 -;chr10 0.13 3.63 3.02E-04 2.91429

some of the QTLs even have SNP and gene in different chromosomes:

cis rs2869546 chr15 78907345 C T SNORA63 chr3 186505088 186505222 + -0.15 -3.82 1.41E-04
cis rs7182583 chr15 78899210 G C SNORA63 chr3 186505088 186505222 + 0.14 3.57 3.77E-04
cis rs9920189 chr15 78666311 G T SNORA63 chr3 186505088 186505222 + 0.14 3.62 3.07E-04

Do you know why does it happen and how to solve it?

andreyshabalin commented 3 years ago

Hi EastsunW,

I'm happy to investigate.

Would you mind sharing your gene/SNP location files you were using? And the genotype and gene expression files (actual values can be replaced with zeros).

My e-mail is: andrey.shabalin@gmail.com

Andrey