caitiecollins / treeWAS

treeWAS: A Phylogenetic Tree-Based Tool for Genome-Wide Association Studies in Microbes
Other
92 stars 18 forks source link

negative association score #76

Open DorothyTamYiLing opened 3 months ago

DorothyTamYiLing commented 3 months ago

Hi,

I am getting negative association scores with my treeWAS run and I am wondering if that is normal? Does negative association mean no association or something else?

Thanks, Dorothy

caitiecollins commented 3 months ago

Hi Dorothy,

Negative association scores are normal: they just describe the direction of association. Typically, for binary phenotypic and genotypic states, a positive association score would indicate that {snps=0 and phen=0} and {snps=1 and phen=1} are associated, whereas a negative association score would indicate that {snps=0 and phen=1} and {snps=1 and phen=0} are associated. The magnitude of the score indicates the extent to which a given snp displays positive over negative association or vice versa.

treeWAS considers the absolute value of the association scores when building the association score distributions and determining the significance threshold, which is why you don't see negative association scores in the plots.

Does that explain what you're seeing or do you have any other concerns? Best, Caitlin.

DorothyTamYiLing commented 3 months ago

Hi Caitlin,

Thank you so much for your reply. I don't have other questions so far, thank you for writing treeWAS again, it is an amazing piece of work.

Thanks, Dorothy

On Fri, 24 May 2024 at 15:53, caitiecollins @.***> wrote:

Hi Dorothy,

Negative association scores are normal: they just describe the direction of association. Typically, for binary phenotypic and genotypic states, a positive association score would indicate that {snps=0 and phen=0} and {snps=1 and phen=1} are associated, whereas a negative association score would indicate that {snps=0 and phen=1} and {snps=1 and phen=0} are associated. The magnitude of the score indicates the extent to which a given snp displays positive over negative association or vice versa.

treeWAS considers the absolute value of the association scores when building the association score distributions and determining the significance threshold, which is why you don't see negative association scores in the plots.

Does that explain what you're seeing or do you have any other concerns? Best, Caitlin.

— Reply to this email directly, view it on GitHub https://github.com/caitiecollins/treeWAS/issues/76#issuecomment-2129733081, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIDXP5M6Q77FMOSZPLBL6B3ZD5H6TAVCNFSM6AAAAABIFWTFGOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMRZG4ZTGMBYGE . You are receiving this because you authored the thread.Message ID: @.***>