[Note: I posted this question on Biostars as well https://www.biostars.org/p/295807/]
I am wondering if anyone here is familiar with this publication [Klingenberg and Meinicke 2017][1]
The authors propose that normalization should be done at the taxonomic level first and then scaled up to the community (global) level when analyzing metatranscriptomic data. They demonstrate that normalizing at the taxonomic level makes differential gene expression analysis more accurate. I am currently working with a metatranscriptome project that contain reads from a number of bacteria, but most if not all of them lack a reference genome, which makes normalization at fine taxonomical levels (e.g. species) almost impossible. So my question is: if I were to apply the method by Kingenberg and Meinicke to my data, would I be able to do so at the genus level? I am not sure if it would be worth the effort.
I could, alternatively, just normalize my data at the global level (i.e. all reads together), as everyone else have been doing before [Klingenberg and Meinicke 2017][1]. However, I believe the authors make a good point and if it is possible to apply their method to my data, I think it would be worth doing so.
[1]: https://peerj.com/articles/3859/
I am wondering if anyone here is familiar with this publication [Klingenberg and Meinicke 2017][1]
The authors propose that normalization should be done at the taxonomic level first and then scaled up to the community (global) level when analyzing metatranscriptomic data. They demonstrate that normalizing at the taxonomic level makes differential gene expression analysis more accurate. I am currently working with a metatranscriptome project that contain reads from a number of bacteria, but most if not all of them lack a reference genome, which makes normalization at fine taxonomical levels (e.g. species) almost impossible. So my question is: if I were to apply the method by Kingenberg and Meinicke to my data, would I be able to do so at the genus level? I am not sure if it would be worth the effort.
I could, alternatively, just normalize my data at the global level (i.e. all reads together), as everyone else have been doing before [Klingenberg and Meinicke 2017][1]. However, I believe the authors make a good point and if it is possible to apply their method to my data, I think it would be worth doing so.
[1]: https://peerj.com/articles/3859/