# Question about GeneMaskScorer and width parameter in gene-level predictions

**URL:** <https://www.alphagenomecommunity.com/t/question-about-genemaskscorer-and-width-parameter-in-gene-level-predictions/681>\
**Category:** Help & Support\
**Created:** [September 30, 2025, 10:01pm UTC](https://www.alphagenomecommunity.com/t/question-about-genemaskscorer-and-width-parameter-in-gene-level-predictions/681 "2025-09-30T22:01:10Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Ilya\_Dyugay](https://yyz2.discourse-cdn.com/watermelon/user_avatar/www.alphagenomecommunity.com/ilya_dyugay/32/580_2.png) [@Ilya\_Dyugay](https://www.alphagenomecommunity.com/u/Ilya_Dyugay)\
**Post date:** [September 30, 2025, 10:01pm UTC](https://www.alphagenomecommunity.com/t/question-about-genemaskscorer-and-width-parameter-in-gene-level-predictions/681/1 "2025-09-30T22:01:10Z")

</div>

Hi,

I am using _score\_interval_ together with _GeneMaskScorer_ to obtain predictions at the gene level. Specifically, I pass gene intervals and aggregate the predictions to get an overall expression estimate for each gene.

What I find unclear is the purpose of the _width_ parameter in _GeneMaskScorer_. The documentation and paper do not fully explain how this parameter is actually used in the scoring process. Could you please explain how _width_ affects the computation of predictions?

Thanks in advance for clarifying!

---

<div class="post-metadata">

**Author:** ![Daniel\_Scott](https://yyz2.discourse-cdn.com/watermelon/user_avatar/www.alphagenomecommunity.com/daniel_scott/32/485_2.png) [@Daniel\_Scott](https://www.alphagenomecommunity.com/u/Daniel_Scott)\
**Post date:** [October 3, 2025, 1:20pm UTC](https://www.alphagenomecommunity.com/t/question-about-genemaskscorer-and-width-parameter-in-gene-level-predictions/681/2 "2025-10-03T13:20:08Z")

</div>

Hello,

Thank you for reaching out. The width in the interval scoring GeneMaskScorer is the width of the target interval to include in the aggregation. Please see this [link](https://github.com/google-deepmind/alphagenome/blob/46e2fe4fbd8b6a03ece00e47851f28e06980ad02/src/alphagenome/models/interval_scorers.py#L82C12-L82C75) for further information.