Request usage of REVEL score

[Marc-Alexander Rauschendorf]

Dear UCSC Team,

 

introduction of REVEL pathogenicity score is a really nice addition and valuable feature.

I have a question regarding the usage of the score. Scores range from 0 to 1 with higher scores predicting a greater likelihood for the variant being disease-causing.

Is there a reasonable threshold in that range for evaluation of benign or pathogenic likelihood? Benign less 0.5 and pathogenic greater or equal 0.5? How should it be applied for variant interpretation?

I did not find any hint regarding a threshold on dbNSFP.

 

Thanks and best regards,

Marc

Dr.​ Marc‑Alexander Rauschendorf Follow Us Project Manager     P:  +496221438512242 F:  +49 6221 43851 2742 M:  +49 163 410 7909 E:  marc.rau...@molecularhealth.com Molecular Health GmbH Kurfürsten‑Anlage 21 69115 Heidelberg www.molecularhealth.com Vorsitzender des Aufsichtsrats: Prof. Dr. Christof Hettich Geschäftsführer: Dr. Friedrich von Bohlen und Halbach (Sprecher der Geschäftsleitung), Dr. Stephan Brock Sitz der Gesellschaft: Heidelberg Handelsregister: Amtsgericht Mannheim - HRB 338037 Confidentiality note: ​The information in this email and any attachment may contain confidential and proprietary information of Molecular Health GmbH, Molecular Health, Inc. and/or its affiliates and may be privileged or otherwise protected from disclosure. If you are not the intended recipient, you are hereby notified that any review, reliance or distribution by others or forwarding without express permission is strictly prohibited and may cause liability. In case you have received this message due to an error in transmission, we kindly ask you to notify the sender immediately and to delete this email and any attachment from your system. ​ ​Disclaimer: The information in this transmission contains privileged and confidential information, including patient information protected by federal and state privacy laws. It is intended only for the use of the person(s) named above. If you are not the intended recipient, you are hereby notified, that any review, dissemination, distribution or duplication of this communication is strictly prohibited. If you are not the intended recipient, please contact the sender by reply mail and destroy all copies of the original message.

[Brian Lee]

Dear Marc,

Thank you for using the UCSC Genome Browser and your question about REVEL score.

To learn more about the REVEL dataset, it may be best to visit their site:
https://sites.google.com/site/revelgenomics
https://sites.google.com/site/revelgenomics/contact?authuser=0

One of our team members did look at one of the publications, https://www.sciencedirect.com/science/article/pii/S0002929716303706?via%3Dihub, and notes that in Figure 5a, https://ars.els-cdn.com/content/image/1-s2.0-S0002929716303706-gr5_lrg.jpg, that around the score of 0.5 the crossover from neutral to disease appears to occur as you describe. This excerpt from the source article is also significant:

    "For example, 75.4% of disease mutations but only 10.9% of neutral variants (and 12.4% of all ESVs) have a REVEL score above 0.5, corresponding to a sensitivity of 0.754 and specificity of 0.891. Selecting a more stringent REVEL score threshold of 0.75 would result in higher specificity but lower sensitivity, with 52.1% of disease mutations, 3.3% of neutral variants, and 4.1% of all ESVs being classified as pathogenic."

Again you may be best served contacting the authors of this REVEL data. You may also be interested to learn we do have a pending paper, Variant Interpretation: UCSC Genome Browser Recommended Track Sets, https://www.authorea.com/users/423801/articles/529043-variant-interpretation-ucsc-genome-browser-recommended-track-sets, that will hopefully be published soon to assist users in applying UCSC Genome Browser resources for clinical purposes.

Thank you again for your inquiry and for using the UCSC Genome Browser. If you have any further public questions, please reply to gen...@soe.ucsc.edu. All messages sent to that address are archived on a publicly accessible forum. If your question includes sensitive data, you may send it instead to genom...@soe.ucsc.edu.

All the best,

--

---
You received this message because you are subscribed to the Google Groups "UCSC Genome Browser Public Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to genome+un...@soe.ucsc.edu.
To view this discussion on the web visit https://groups.google.com/a/soe.ucsc.edu/d/msgid/genome/PAXPR04MB81916E8389DEBCDAB0B2B730F2CC9%40PAXPR04MB8191.eurprd04.prod.outlook.com.

[Marc-Alexander Rauschendorf]

Dear Brian,

 

many thanks for your swift and comprehensive reply.

And thanks for pointing me to the figure. I was going to fast through the paper screening for cut-off as a keyword and I missed the explanation in the graph. Also, in the dbNSFP release notes they did not mention any cut-off for REVEL either though they give one for most of the other prediction scores.

 

We just finished the ESHG 2021 meeting and so many colleagues talked about the REVEL score for variant prioritization and that it was now also available in the UCSC browser. It is really a nice addition.

 

Best,

Marc

 

 

Von: Brian Lee <bria...@soe.ucsc.edu>
Gesendet: Wednesday, September 1, 2021 01:54
An: Marc-Alexander Rauschendorf <marc.rau...@molecularhealth.com>
Cc: gen...@soe.ucsc.edu
Betreff: Re: [genome] Request usage of REVEL score

 

[CAUTION: This email originated from outside of Molecular Health. Do not click links or open attachments unless you recognize the sender and are expecting the attachment or link.]

[Robert Kuhn]

Dear Marc,

It's great to hear that people at ESHG noticed and mentioned the Browser talk and

REVEL score.  It's so difficult to know in the virtual format if anyone is listening or

interested.

> did not mention any cut-off for REVEL either

As you can tell in the graph that Brian mentioned, the REVEL is a sliding scale and no

real cutoff exists.  It's a moving probability.  It's probably no accident that the authors do

not define a cutoff score.

best wishes and good luck in your work,

   --b0b kuhn

   ucsc genome bioinformatics group

On Thu, Sep 2, 2021 at 9:07 AM Marc-Alexander Rauschendorf <marc.rau...@molecularhealth.com> wrote:

Dear Brian,

 

many thanks for your swift and comprehensive reply.

And thanks for pointing me to the figure. I was going too fast through the paper screening for cut-off as a keyword and I missed the explanation in the graph. Also, in the dbNSFP release notes they did not mention any cut-off for REVEL either though they give one for most of the other prediction scores.

To view this discussion on the web visit https://groups.google.com/a/soe.ucsc.edu/d/msgid/genome/PAXPR04MB81914E347CFD8F7EFA4D23EFF2CE9%40PAXPR04MB8191.eurprd04.prod.outlook.com.

[Marc-Alexander Rauschendorf]

Dear Robert,

 

many thanks for your reply.

Yes, the sliding scale with likelihood of higher scores for variants being more disease-causing vs. a defined cut-off is kind of a tricky thing.

We want to use it for variant prioritization and filtering. In this case a cut-off is easier to use. We will start with a cut-off at 0.5 as mentioned before. But we will re-evaluate after some time to see if this is the best approach.

 

Thanks for bringing the scores into UCSC browser.

 

Best regards,

Marc