search

gagneurlab / drop

Pipeline to find aberrant events in RNA-Seq data, useful for diagnosis of rare disorders
MIT License
133 stars 44 forks source link

OUTRIDER > Overexpression of TBR1 not found #583

Open juugeebee opened 1 day ago

juugeebee commented 1 day ago

Hi,

The attached ppt presents the case of overexpression of TBR1 due to a PSMD14/TBR1 fusion on patient 6624CY000185. We ran OUTRIDER via DROP and this variant was not found.

You can find in that link the raw OUTRIDER outputs files and the DROP config file : https://dispose.aphp.fr/userportal/?v=4.5.6#/shared/public/8c_Zd35P2Xrjz9lo/86866af4-a635-48d1-ad22-adcdd85abb95

Would you have any clues that could explain why the tool does not find the variant? PSMD14-TBR1.pptx

Thanks in advance for your help! Julie Bogoin

vyepez88 commented 1 day ago

Hi, I cannot access the link, it says: "Could not find shared item. It has been either moved or deleted, or the link has expired.". But, from what I can see in the IGV screenshots, the TBR1 is usually not expressed in your tissue, so it is filtered out. OUTRIDER can only model genes that are sufficiently expressed. Maybe you can find it via splicing analysis.

vyepez88 commented 1 day ago

Indeed, it seems the gene is only expressed in brain: https://www.gtexportal.org/home/gene/TBR1

juugeebee commented 1 day ago

Thanks for your answer! I updated the link : https://dispose.aphp.fr/userportal/?v=4.5.6#/shared/public/8c_Zd35P2Xrjz9lo/86866af4-a635-48d1-ad22-adcdd85abb95

juugeebee commented 1 day ago

For this run, we add at least 1 control, we know it has an overexpression of TBR1. Won't it be enough to overpass filtering?

vyepez88 commented 1 day ago

Hi, it seems that your samples are duplicated {id} and {id}_ref seem to be the same. Is this intended? Having only 1 or 2 samples passing the filtering is not enough as the default filter of OUTRIDER is an FPKM > 1 in at least 5% of the samples. I plotted the raw data and you can see that only the sample that want (and it's replicate) are overexpressed in the whole cohort: image

juugeebee commented 1 day ago

It is indeed intended. It was a way we thought to overpass the filtering...

vyepez88 commented 1 day ago

It's not a good idea as it affects the modelling assumptions of OUTRIDER. You can change the filtering parameters fpkmCutoff and percentile of the filterExpression function. However, this can lead to keeping unwanted genes. I recommend simply taking the plot as confirmation that it is an extreme change as no one else in the cohort had it.