# Differential analysis on gene expression data

**URL:** <https://forum.depmap.org/t/differential-analysis-on-gene-expression-data/3148>\
**Category:** Q&A\
**Tags:** omics, data\
**Created:** [March 20, 2024, 1:08pm UTC](https://forum.depmap.org/t/differential-analysis-on-gene-expression-data/3148 "2024-03-20T13:08:13Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![Changyuan\_Hu](https://yyz1.discourse-cdn.com/flex035/user_avatar/forum.depmap.org/changyuan_hu/32/1079_2.png) [@Changyuan\_Hu](https://forum.depmap.org/u/Changyuan_Hu)\
**Post date:** [March 20, 2024, 1:08pm UTC](https://forum.depmap.org/t/differential-analysis-on-gene-expression-data/3148/1 "2024-03-20T13:08:13Z")

</div>

Hi  
I am wondering if there is any pipeline for differential analysis on log2(TPM+1) data? Can I simplify use limma or other R package recommended?

Thanks ahead for your comments.

---

<div class="post-metadata">

**Author:** ![dllahr](https://yyz1.discourse-cdn.com/flex035/user_avatar/forum.depmap.org/dllahr/32/23_2.png) [@dllahr](https://forum.depmap.org/u/dllahr)\
**Post date:** [March 21, 2024, 12:50pm UTC](https://forum.depmap.org/t/differential-analysis-on-gene-expression-data/3148/2 "2024-03-21T12:50:28Z")

</div>

Hello - I like to use voom-limma for RNA-seq analysis, but I would recommend instead of using log2(TPM+1) data that use the read count data (also available for download), and follow standard practices/workflow of doing library normalization etc.

---

<div class="post-metadata">

**Author:** ![Changyuan\_Hu](https://yyz1.discourse-cdn.com/flex035/user_avatar/forum.depmap.org/changyuan_hu/32/1079_2.png) [@Changyuan\_Hu](https://forum.depmap.org/u/Changyuan_Hu)\
**Post date:** [March 24, 2024, 1:18pm UTC](https://forum.depmap.org/t/differential-analysis-on-gene-expression-data/3148/3 "2024-03-24T13:18:48Z")

</div>

Hi dllahr,

Thanks for quick response, but do you have more detailed workflow for differential analysis begin with the read count data?  
As far as I googled, we can download ExpectedCountProfile, AllGenesEffectiveLengthProfile, and AllGenesTPMLogp1Profile, and import these three matrix to package tximport for DESeq2 analysis. is it right?

---

<div class="post-metadata">

**Author:** ![dllahr](https://yyz1.discourse-cdn.com/flex035/user_avatar/forum.depmap.org/dllahr/32/23_2.png) [@dllahr](https://forum.depmap.org/u/dllahr)\
**Post date:** [March 24, 2024, 2:00pm UTC](https://forum.depmap.org/t/differential-analysis-on-gene-expression-data/3148/4 "2024-03-24T14:00:04Z")

</div>

We have our own internal pipeline (not published unfortunately) but these days I tend to check next flow / nf-core for great starting points for standard pipelines and analyses. The one for differential gene expression is called “differential abundance” because it can be applied more generally than RNA-seq but it looks pretty good to me:

> **[differentialabundance: Introduction](https://nf-co.re/differentialabundance/1.4.0)**
>
> Differential abundance analysis for feature/ observation matrices from platforms such as RNA-seq

---

<div class="post-metadata">

**Author:** ![Changyuan\_Hu](https://yyz1.discourse-cdn.com/flex035/user_avatar/forum.depmap.org/changyuan_hu/32/1079_2.png) [@Changyuan\_Hu](https://forum.depmap.org/u/Changyuan_Hu)\
**Post date:** [March 25, 2024, 6:00am UTC](https://forum.depmap.org/t/differential-analysis-on-gene-expression-data/3148/5 "2024-03-25T06:00:14Z")

</div>

That is so nice of you, Dave. Much appreciated for you kindly sharing.