# P-values and q-values are identical for custom analysis of drug sensitivity AUC?

**URL:** <https://forum.depmap.org/t/p-values-and-q-values-are-identical-for-custom-analysis-of-drug-sensitivity-auc/3145>\
**Category:** Report an Issue\
**Created:** [March 19, 2024, 6:39pm UTC](https://forum.depmap.org/t/p-values-and-q-values-are-identical-for-custom-analysis-of-drug-sensitivity-auc/3145 "2024-03-19T18:39:11Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![David\_Lu](https://yyz1.discourse-cdn.com/flex035/user_avatar/forum.depmap.org/david_lu/32/1076_2.png) [@David\_Lu](https://forum.depmap.org/u/David_Lu)\
**Post date:** [March 19, 2024, 6:39pm UTC](https://forum.depmap.org/t/p-values-and-q-values-are-identical-for-custom-analysis-of-drug-sensitivity-auc/3145/1 "2024-03-19T18:39:11Z")

</div>

I am using the custom analysis function to output Drug sensitivity AUC (Prism Repurposing Secondary Screen) correlations and q-values (See attached image).

 ![Screenshot 2024-03-18 230522](https://canada1.discourse-cdn.com/flex035/uploads/depmap/original/2X/2/2f01e60b4430f81feec1468263a57923d19c429c.png)

However, almost of the calculated p-values and q-values for each drug are identical to each other (see attached image). Shouldn’t the q-value generally be lower than the p-value?  
 ![Screenshot 2024-03-18 231121](https://canada1.discourse-cdn.com/flex035/uploads/depmap/original/2X/4/4d8736b00fa107452c1e16d8be60457d210e9952.png)

Thank you for your help!

---

<div class="post-metadata">

**Author:** ![lorcai](https://avatars.discourse-cdn.com/v4/letter/l/a88e57/32.png) [@lorcai](https://forum.depmap.org/u/lorcai)\
**Post date:** [March 20, 2024, 5:00pm UTC](https://forum.depmap.org/t/p-values-and-q-values-are-identical-for-custom-analysis-of-drug-sensitivity-auc/3145/2 "2024-03-20T17:00:07Z")

</div>

Hi,

Just to note, I posted a similar problem recently:  
[Repeated uniprot accesions and weird qvalues of proteomic data in the analysis tool](https://forum.depmap.org/t/repeated-uniprot-accesions-and-weird-qvalues-of-proteomic-data-in-the-analysis-tool/3137)

It doesn’t show in the screencap but I’m guessing that numCellLines was variable across compounds. In my case (running pearson correlation for MetMap500 vs Proteomic data) it seemed that when values were present across all queried cell lines (numCellLines=all cell lines queried), the pvalue was adjusted (in some way that I haven’t been able to replicate though), but for the rest of the cell lines it’s often just pvalue=qvalue.

I think there’s is a problem with the pvalue adjustment when some of the queried cell lines have no values…