# Zero value with scvi normalisation

**URL:** <https://discourse.scverse.org/t/zero-value-with-scvi-normalisation/3352>\
**Category:** scvi-tools\
**Tags:** scvi\
**Created:** [November 20, 2024, 11:28am UTC](https://discourse.scverse.org/t/zero-value-with-scvi-normalisation/3352 "2024-11-20T11:28:21Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![LioLnr](https://yyz1.discourse-cdn.com/flex035/user_avatar/discourse.scverse.org/liolnr/32/1108_2.png) [@LioLnr](https://discourse.scverse.org/u/LioLnr)\
**Post date:** [November 20, 2024, 11:28am UTC](https://discourse.scverse.org/t/zero-value-with-scvi-normalisation/3352/1 "2024-11-20T11:28:21Z")

</div>

Hello,

I did a normalisation with scvi-tools. Now I want to use the dotplot function of scanpy.  
Unfortunatly, the normalisation seems to have modified all the zero value to non zero. Thus when using the dotplot function I do not have anymore information about the % of cells expressing the genes. There si no more zero value so 100% of cells express each genes.

How would you deal with that ?

---

<div class="post-metadata">

**Author:** ![cane11](https://yyz1.discourse-cdn.com/flex035/user_avatar/discourse.scverse.org/cane11/32/241_2.png) [@cane11](https://discourse.scverse.org/u/cane11)\
**Post date:** [November 27, 2024, 6:39am UTC](https://discourse.scverse.org/t/zero-value-with-scvi-normalisation/3352/2 "2024-11-27T06:39:48Z")

</div>

Hi, you can use the expression\_cutoff argument in scanpy.pl.dotplot and adjust it or you can use the [posterior\_predictive\_sample](https://docs.scvi-tools.org/en/latest/api/reference/scvi.model.SCANVI.html#scvi.model.SCANVI.posterior_predictive_sample) function to actually generate count data. However, this function does not support projecting to batches.