# Details on methodology of “Two Class Comparison”

**URL:** <https://forum.depmap.org/t/details-on-methodology-of-two-class-comparison/211>\
**Category:** Feedback\
**Created:** [September 20, 2020, 10:09pm UTC](https://forum.depmap.org/t/details-on-methodology-of-two-class-comparison/211 "2020-09-20T22:09:46Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![shovikb](https://avatars.discourse-cdn.com/v4/letter/s/2bfe46/32.png) [@shovikb](https://forum.depmap.org/u/shovikb)\
**Post date:** [September 20, 2020, 10:09pm UTC](https://forum.depmap.org/t/details-on-methodology-of-two-class-comparison/211/1 "2020-09-20T22:09:47Z")

</div>

Would it be possible to include some details on the two class comparison custom analysis method? I assume it’s a two sample hypothesis test with some form of multiple hypothesis correction, but the exact method that is used would be useful for reporting and reproducibility.

It’s a great tool overall though, really has sped up the process of working through a lot of datasets quickly.

---

<div class="post-metadata">

**Author:** ![Mustafa\_Kocak](https://yyz1.discourse-cdn.com/flex035/user_avatar/forum.depmap.org/mustafa_kocak/32/11_2.png) [@Mustafa\_Kocak](https://forum.depmap.org/u/Mustafa_Kocak)\
**Post date:** [October 9, 2020, 1:15pm UTC](https://forum.depmap.org/t/details-on-methodology-of-two-class-comparison/211/2 "2020-10-09T13:15:46Z")

</div>

Hi Shovik,

As you guessed correctly , the two class comparison simply consists of a simple linear hypothesis test followed up a multiple hypothesis correction step.

In particular, for each feature/column of the selected dataset we are fitting a simple linear regression model to the chosen phenotype of interest. The estimated regression coefficient and its standard error then fed into the adaptive shrinkage method described in [https://doi.org/10.1093/biostatistics/kxw041](https://doi.org/10.1093/biostatistics/kxw041) to obtain moderated effect sizes (posterior mean estimates) and corresponding q-values (FDR). For the binary features, this methodology is being roughly equivalent (the only difference is the shrinkage step) to using a t-test with a pooled variance estimate.

Also, for the sake of reproducibility we keep the code used in this analysis in this github repo: [https://github.com/broadinstitute/cdsr\_models/blob/master/R/linear\_association.R](https://github.com/broadinstitute/cdsr_models/blob/master/R/linear_association.R)

Warmly,  
Mustafa

---

<div class="post-metadata">

**Author:** ![aobla](https://avatars.discourse-cdn.com/v4/letter/a/9fc348/32.png) [@aobla](https://forum.depmap.org/u/aobla)\
**Post date:** [August 19, 2024, 1:07pm UTC](https://forum.depmap.org/t/details-on-methodology-of-two-class-comparison/211/3 "2024-08-19T13:07:50Z")

</div>

I would like to confirm if the function “run\_lm\_stats\_limma” in [cdsr\_models/R/linear\_association.R at master · broadinstitute/cdsr\_models · GitHub](https://github.com/broadinstitute/cdsr_models/blob/master/R/linear_association.R) is still being used for two class comparison found in the latest version of Data Explorer ?

---

<div class="post-metadata">

**Author:** ![MarkoFlow](https://yyz1.discourse-cdn.com/flex035/user_avatar/forum.depmap.org/markoflow/32/1687_2.png) [@MarkoFlow](https://forum.depmap.org/u/MarkoFlow)\
**Post date:** [May 11, 2026, 12:29am UTC](https://forum.depmap.org/t/details-on-methodology-of-two-class-comparison/211/4 "2026-05-11T00:29:23Z")

</div>

Thank you, Mustafa. This clarifies the methodology and is helpful for interpreting the results.