Question: Relation between Log2 Fold change and Log2 TPM in RNA-seq data
1
3 months ago by
saigro1610
saigro1610 wrote:

Hello,

I have one basic question, I would like to know the relationship between Log2 Fold change and Log2 TPM in RNA-seq data. Specifically, One particular gene A Log2 TPM is 9 and its Log2 Foldchange 3 fold. I am not able to understand how this gene TPM is higher and foldchange is low comparatively. Can anybody explain?

Thanks, Ni

rna-seq R • 413 views
modified 3 months ago by grant.hovhannisyan1.5k • written 3 months ago by saigro1610
3
3 months ago by
ATpoint15k
Germany
ATpoint15k wrote:

log2(TPM) is simply the log2 of the Transcript Count Per Million. TPM is a normalization technique (but not a good one) to scale the read count per gene/transcript towards the total read count of the sequencing run in order to compensate for different sequencing depths. See this video for details. The Log2FC is a measure of how much the expression of a gene changes between two conditions.

3
3 months ago by
i.sudbery4.3k
Sheffield, UK
i.sudbery4.3k wrote:

TPM (Transcripts Per Million) refers to how much RNA is present in a sample. For example, a Log2 TPM of 9 means that for every million transcripts in your sample, 2^9 of them are from gene A. It is the expression level of gene A in a sample.

Log2Foldchange describes how one sample is different from another. In this case it is saying that the expression level in Condition 2 is 6 (2^3) times as high as it is in Condition 1.

So if the log2 expression level in condition 1 is 9 TPM, and there is a log2foldchange of 3 between condition 1 and condition 2, then the log2 expression level in condition 2 will be 12 TPM (2^9 * 2^3).

Thank for the clear explanation, I have one more question how should we explain when log2TPM of a particular gene is around 5 and its log2 foldchange is -1.15?

That means it is downregulated in the condition you compare it with, so fewer counts and therefore a negative fold change.

Just so. Thus, for a log2foldchange of -1.15, this means that the expression in the treament condition is 2^-1.15 = 0.45 of the expression in the control condition. That is the treatment expression is 1/0.45=2.2 fold lower.

Thank you for the reply, regarding Log2 foldchange -1.15 I got it but I am not able to understand how does it relate to Log2 TPM is in positive value 5.

It doesn't relate to this.

A gene cannot have one log2TPM in two conditions, nor a log2 fold change in two conditions.

If you have two condidtions, you have two different values from log2TPM, one in each condition.

If, for example your log2 TPM is 5 in the control condition, and the log2Fold change between control and treatment is -1.15, then the expression level in the treatment condition would be 3.85.

2
3 months ago by
grant.hovhannisyan1.5k wrote:

Don't compare these two units - its not apples to apples comparison. TMP is a unit of expression. Fold Change is a unit showing how much TPM has changed in one condition compared to another condition.