# Autoencoder gene expression reconstruction accuracy

**URL:** <https://discourse.scverse.org/t/autoencoder-gene-expression-reconstruction-accuracy/470>\
**Category:** scvi-tools\
**Created:** [May 19, 2022, 9:45pm UTC](https://discourse.scverse.org/t/autoencoder-gene-expression-reconstruction-accuracy/470 "2022-05-19T21:45:35Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![rwollman](https://yyz1.discourse-cdn.com/flex035/user_avatar/discourse.scverse.org/rwollman/32/237_2.png) [@rwollman](https://discourse.scverse.org/u/rwollman)\
**Post date:** [May 19, 2022, 9:45pm UTC](https://discourse.scverse.org/t/autoencoder-gene-expression-reconstruction-accuracy/470/1 "2022-05-19T21:45:35Z")

</div>

Hi scvi folks,

We had a discussion in the lab, and we’re hoping some scvi-verse expert could chime in.

A naive question here: what reconstruction accuracy should one expect from scvi tools in terms of total variance explained? ballpark value, is it 1%, 10%, 90%?

Where reconstruction accuracy is : mean((X-X’)^2)/mean((X-mean(X))^2) where X is cell x gene input matrix to scvi and X’ is the output of the autoencoder.

Thanks,  
Roy

---

<div class="post-metadata">

**Author:** ![PierreBoyeau](https://yyz1.discourse-cdn.com/flex035/user_avatar/discourse.scverse.org/pierreboyeau/32/29_2.png) [@PierreBoyeau](https://discourse.scverse.org/u/PierreBoyeau)\
**Post date:** [June 1, 2022, 7:00pm UTC](https://discourse.scverse.org/t/autoencoder-gene-expression-reconstruction-accuracy/470/2 "2022-06-01T19:00:59Z")

</div>

Thanks for reaching out! We are working on designing a model criticism module for scvi-tools, and metrics like the one you describe could be a good addition. Do you have references describing the use of this metric?