Differences between codon usage bias vs. differential transcript usage (DTU)

Question

[nickpark1997]

I'm wondering what the key biological differences are between codon usage bias (how frequently different codons are used) and DTU (differential transcript usage), and how those differences inform the conclusions we can draw from these two techniques. For reference, this paper describes DTU as follows:

Differential transcript usage (DTU) is a special case of alternative splicing in which a gene’s relative transcript abundance (RTA) profile, the proportion of total gene-level expression that arises from each of the gene’s transcript isoforms, differs between conditions (Soneson et al., 2016). [...] An example of DTU may be illustrated by the scenario where the total gene expression is the same between conditions A and B, but transcript isoform 1 is the predominantly expressed transcript in condition A while transcript isoform 2 is the predominantly expressed transcript in condition B.

Specifically, I'm trying to understand what different conclusions we could draw from analysis of a codon usage frequency dataset vs. DTU. My current understanding is that codons tell you which proteins are made, and thus are able to illustrate larger scale differences between organisms of different species, while DTU has to do with measuring mutations through alternative splicing before proteins are made and is primarily used to examine special conditions within the same species. An intra- vs. inter-species distinction.

However, I suspect that my understanding has major holes and that these two concepts can be compared in a more nuanced way than a simple intra- vs. inter-species distinction.

Any help to clarify/expand on the differences would be much appreciated, thank you!

Answer 1

[i.sudbery]

Codon usage bias refers to the phenomenon that there are several codons that can code for the same amino acid. So, for example, the codons GCT, GCC GCA and GCG all encode the amino acid Alanine. However, despite the fact that any of these four base sequences encode alanine, genomes tend to have a preference, or bias to use some more frequently than others. For example, in the E. coli genome 33% of alanines are coded for by GCG triplets and only 18% by GCT triplets.

Codon usage bias is most commonly computed across a whole genome. However, you can have difference in codon usage bias between two different cell types in the same organism, or even between two different conditions for the same cell type. This would arise if the two cell types in question expressed different sets of genes that had different preferences.

Codon usage bias is particularly important in synthetic biology and biological engineering because it is important for efficient expression that the sequence of any gene you introduce matches the natural codon usage bias of the cells you are introducing it into.

Differential transcript usage reffers to alternative splicing and in particular the comparison between two cell types or the same cell type under two conditions.

Consider a gene with three exons:

|>>>>>>>>>>>>>>>>>>|------------|>>>>>>>>>>>>>>>>>>>>>|------------|>>>>>>>>>>>>>>>>|

If we have two cell types and in the first cell type the first two exons are used:

|>>>>>>>>>>>>>>>>>>|------------|>>>>>>>>>>>>>>>>>>>>>>|

and in the second cell type the first and third exons are used:

|>>>>>>>>>>>>>>>>>>|-----------------------------------------------|>>>>>>>>>>>>>>>>|

We would say that there is differential transcript usage.

This could lead to different proteins being produced or the two transcripts might contain different regulatory sites, or one might fail to make a functional protein entirely. Thus it is important in understanding how the protein complement of a cell is different between cell types or conditions.

Answer 2

[Mensur Dlakic]

Codon usage bias is a global measure of how codons (or tRNAs, if you will) are used by a particular organism during translation. It is more productive and easier to regulate and make a lot of 20-30 different tRNAs than to make a little bit of all 61 tRNAs (in an ideal case, as some tRNAs may not even exist in the genome).

Differential transcript usage is a measure of how individual transcript variants are used differently in distinct tissues, or maybe in the same tissue but in different developmental stages.