# Unable to recreate substitution mutations in AlphaGenome

**URL:** <https://www.alphagenomecommunity.com/t/unable-to-recreate-substitution-mutations-in-alphagenome/706>\
**Category:** Help & Support\
**Tags:** testing\
**Created:** [November 10, 2025, 3:00am UTC](https://www.alphagenomecommunity.com/t/unable-to-recreate-substitution-mutations-in-alphagenome/706 "2025-11-10T03:00:14Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Tee\_Dee](https://yyz2.discourse-cdn.com/watermelon/user_avatar/www.alphagenomecommunity.com/tee_dee/32/569_2.png) [@Tee\_Dee](https://www.alphagenomecommunity.com/u/Tee_Dee)\
**Post date:** [November 10, 2025, 3:00am UTC](https://www.alphagenomecommunity.com/t/unable-to-recreate-substitution-mutations-in-alphagenome/706/1 "2025-11-10T03:00:14Z")

</div>

Hello,

I am interested in running some test cases for AlphaGenome to get a sense of how it works.

I want to see the change in RNA expression in Lung, when a substitution causes a stop codon to be produced.

To see this, I wanted to recreate a mutation in CFTR namely the substitution of G to T, causing Glycine to change to a stop codon. GGA → UGA

I have taken this data from a credible source:

> **[NM\_000492.4(CFTR):c.1624G\>T (p.Gly542Ter) AND Cystic fibrosis - ClinVar -...](https://www.ncbi.nlm.nih.gov/clinvar/RCV000007535/)**
>
> ClinVar archives and aggregates information about relationships among variation and human health.

I used this data: \* NC\_000007.14:g.117587778G\>T  
as AlphaGenome deals with the genomic sequence.

This is the code I used:

 ![image](https://canada1.discourse-cdn.com/watermelon/original/1X/cf928e3c9d0a24e9e3105ed9fa52b66d6099b47e.png)

 ![image](https://canada1.discourse-cdn.com/watermelon/original/1X/55908914ae56f251842327032d674a46c38e7d36.png)

However, there seems to be no difference in the Reference vs Alternative Plots.

I even cross checked the nucleotide number with the UCSC data.  
[https://genome.ucsc.edu/cgi-bin/hgTracks?db=hg38&lastVirtModeType=default&lastVirtModeExtraState=&virtModeType=default&virtMode=0&nonVirtPosition=&position=chr7%3A117587763-117587826&hgsid=3314559807\_i8e2GbbTivHOglXUsMheuNzmHqRS](https://genome.ucsc.edu/cgi-bin/hgTracks?db=hg38&lastVirtModeType=default&lastVirtModeExtraState=&virtModeType=default&virtMode=0&nonVirtPosition=&position=chr7%3A117587763%2D117587826&hgsid=3314559807_i8e2GbbTivHOglXUsMheuNzmHqRS)

I have been struggling with this issue for quite some time now, and would greatly appreciate some feedback.

Is this an issue with AlphaGenome or an issue with my code?

Thank you

---

<div class="post-metadata">

**Author:** ![Nicolene\_Pillay](https://yyz2.discourse-cdn.com/watermelon/user_avatar/www.alphagenomecommunity.com/nicolene_pillay/32/655_2.png) [@Nicolene\_Pillay](https://www.alphagenomecommunity.com/u/Nicolene_Pillay)\
**Post date:** [November 11, 2025, 10:07am UTC](https://www.alphagenomecommunity.com/t/unable-to-recreate-substitution-mutations-in-alphagenome/706/2 "2025-11-11T10:07:05Z")

</div>

Hello,

Thanks for reaching out. The code looks fine, and it looks like the variant input is correct, at least it matches the variant ID provided in the NCBI ClinVar record provided. We cannot comment on the robustness of the assumption that the variant is pathogenic.

We recommend also looking at the variant scores for this variant rather than just the REF/ALT visualisations, as sometimes subtle effects can be hard to observe in the visualisations (see the [Tutorial Scoring and Visualising a single variant](https://www.alphagenomedocs.com/colabs/variant_scoring_ui.html) for how to do this).

While we extensively measured performance of AlphaGenome across a variety of modalities and use-cases (see the preprint), there is no guarantee that AlphaGenome predictions will match expectations from known, or putative biology in every case.

---

<div class="post-metadata">

**Author:** ![Tee\_Dee](https://yyz2.discourse-cdn.com/watermelon/user_avatar/www.alphagenomecommunity.com/tee_dee/32/569_2.png) [@Tee\_Dee](https://www.alphagenomecommunity.com/u/Tee_Dee)\
**Post date:** [November 12, 2025, 1:49am UTC](https://www.alphagenomecommunity.com/t/unable-to-recreate-substitution-mutations-in-alphagenome/706/3 "2025-11-12T01:49:52Z")

</div>

Hello, Thank you very much for your reply. I tried to recreate the code as in the tutorial, just changing the variant for CFTR, but end up getting the same plot as I had gotten above, with no difference in the reference and alternative sequence, as indicated in the image below:

 ![image](https://canada1.discourse-cdn.com/watermelon/original/1X/b181d1d3b080a3d28155d06e10eb2c3030329f0c.png)

Does this mean that this is a test case that AlphaGenome is not able to recreate? I thought that because it is a simple substitution to generate a stop codon, we will be able to see a difference, however small.

Can I assume that AlphaGenome cannot give me a correct prediction for this test case?