Seqweaver DIS vs RBP dysregulation

[goutha...@gmail.com]

Hi All,

In the recent NG paper  "Genome-wide landscape of RNA-binding protein target site dysregulation reveals a major impact on psychiatric disorder risk" the "RBP dysregulation"  is used instead of Seqweaver DIS.

I am wondering if its the mean of absolute predicted probability differences of all 232 RBPs for each variant ?

Thanks,

Goutham A

[Christopher Park]

Hi Goutham,

Seqweaver DIS is a weighted score version of RBP scores via logistic regression outlined in Zhou et al 2019 Nature Genetics.

In the recent study, we used the individual RBP dysregulation scores, since we wanted to identify RBPs associated with psychiatric disease risk.

Best,

Chris

[goutha...@gmail.com]

Hi Chris,

Thanks for this info. Sorry, i haven't got back as I was busy with the thesis. I am able to get DIS after running Seqweaver through 'code_asd_dnarna_v2'. Wondering if its possible to get "RBP-dysregulation" or is it same as the variant-level score that is used to compute DIS ? 

On separate note, may I know what does a negative DIS mean in Seqweaver ? Can I use abs(DIS) to look for the effect of a variant ?

Thanks,

Goutham A

[Christopher Park]

Hi Goutham,

DIS can be considered a weighted average of "RBP-dysregulation" across 232 RBPs. You can look at each RBP model in isolation if you want (the results will be in the "seqweaver_human/*diffs.h5") subdirectories, really depends on what your question is. A the variant-level score can be looked at via DIS or each RBP feature in isolation if you want to check which RBP might be disrupted but this would require multiple hypothesis correction.

For DIS these scores are in Z-scores with mean, sd from the 1000Genome variants used as background. So a negative DIS just means that its ranked low in the distribution using abs(DIS) wouldn't really make sense here. We usually only use absolute values when looking at each individual factor (e.g. TF or RBP) disruption via a variant i.e. ABS( Prob_binding(Alt allele) - Prob_binding(Ref allele) ) .

Best,

Chris

[goutha...@gmail.com]

Hi Chris,

Thanks for the explanation. 

-

Goutham A