relationship between coverage and expression level in RNAseq

Question

[shl198]

I wonder what is the relationship between the depth and gene expression level in RNAseq?

For example, let's say I have RNAseq with 10M reads for one sample. If I use the same machine and generate 20M reads for the same sample, then except for the new detected genes, will the expression profile for genes expressed in both case be similar? In other words, if I plot the density VS log2(count) for them, will they have the similar shape?

Answer 1

[Diego D.]

Yes, if you use the same method to normalize the counts, you will get similar MA plots (in theory). At greater coverage, maybe you will see more dots, because there will be more genes detected.

The reads generated during the sequencing should be randomly distributed, except if the process runs into a problem. After normalization, you should have similar values for genes that were detected in both cases.

More coverage is useful to detect low expressed genes.

You will see significant differences in MA plot if you change the number of biological replicates. With more replicates, there will be more differentially expressed genes.

Answer 2

[Gary]

Two related papers for your information.

Wang, Y., Ghaffari, N., Johnson, C.D., Braga-Neto, U.M., Wang, H., Chen, R. & Zhou, H. (2011). Evaluation of the coverage and depth of transcriptome by RNA-Seq in chickens. BMC Bioinformatics 12, S5.

Sims, D., Sudbery, I., Ilott, N.E., Heger, A. & Ponting, C.P. (2014). Sequencing depth and coverage: key considerations in genomic analyses. Nat. Rev. Genet. 15, 121-132.