# Data download from Atlas expression dataset

**URL:** <https://community.opentargets.org/t/data-download-from-atlas-expression-dataset/1011>\
**Category:** Data downloads\
**Tags:** ot-platform, expression-atlas\
**Created:** [27 March 2023 12:49 UTC](https://community.opentargets.org/t/data-download-from-atlas-expression-dataset/1011 "2023-03-27T12:49:48Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![EAbad](https://avatars.discourse-cdn.com/v4/letter/e/6f9a4e/32.png) [@EAbad](https://community.opentargets.org/u/EAbad)\
**Post date:** [27 March 2023 12:49 UTC](https://community.opentargets.org/t/data-download-from-atlas-expression-dataset/1011/1 "2023-03-27T12:49:48Z")

</div>

Hi, I’ve been trying to download Atlas Expression data, fold-change, p-values and pathways involved in each target in the Atlas expression dataset. I use Python code described in OT web page for accessing to data dowloads ([Download datasets - Open Targets Platform Documentation](https://platform-docs.opentargets.org/data-access/datasets)). So far it worked for i.e. targetID, fold-changes, disease, or even a score for evidence, but it does not work with p-values or pathways value columns, they appear as null values. All these features are in the same level of hierarchy in the schema as targetId or fold-changes. Has anyone faced this same problem, could you help me please? Many thanks in advance.

---

<div class="post-metadata">

**Author:** ![hcornu](https://dub1.discourse-cdn.com/flex017/user_avatar/community.opentargets.org/hcornu/32/716_2.png) [@hcornu](https://community.opentargets.org/u/hcornu)\
**Post date:** [12 April 2023 14:18 UTC](https://community.opentargets.org/t/data-download-from-atlas-expression-dataset/1011/2 "2023-04-12T14:18:33Z")

</div>

Hi @EAbad, and welcome to the Open Targets Community! 🎉

You can see the information we have available for differential expression in the Expression Atlas Table, for example: [Open Targets Platform](https://platform.opentargets.org/evidence/ENSG00000143631/EFO_0000274)

What additional information are you trying to get?

---

<div class="post-metadata">

**Author:** ![EAbad](https://avatars.discourse-cdn.com/v4/letter/e/6f9a4e/32.png) [@EAbad](https://community.opentargets.org/u/EAbad)\
**Post date:** [12 April 2023 16:14 UTC](https://community.opentargets.org/t/data-download-from-atlas-expression-dataset/1011/3 "2023-04-12T16:14:01Z")

</div>

Dear Hellen, thank you so much for your attention. Right, certainly, I found expression data in OT platform inside the Atlas expression dataset, but I had issues trying to find the p-value when using parquet files in Python (using the example of how to make queries with pyspark). Finally, a colleague has found that the p-value is not in the log-pvalue columns labeled in the dataset (in these ones we found all null values) but in another column labeled as ‘resourceScore’, that was quite confusing. Also, I got errors when downloading baseline expression data because with parquet files seemed to be not so straightforward and finally we’re using json format. Altogether it seems good so far. Thanks a lot for your message.