Meta transcriptomics without test/control samples.

Question

[SKY]

Hello!

I need to find out microbes that are 'ACTIVE' in a particular environment. Lets call this as treated environment.

My question is simple, do i need to compare it with another Control/sterile environment to determine 'ACTIVE' members of treated environment.

In that case how can one find genes that are Differentially Expressed?

I remain confused. Please help.

Thanks in advance! Akash D

[andres.firrincieli]

The identification of Differentially expressed (Abundant?) transcripts in metatranscriptomic data can be tricky because the abundance of transcripts is affected both by the species abundance and gene abundance. To account for variations caused by gene and species abundances you would need the metagenome. If you only have the metatranscriptomic data then you can try the normalization models M1, M2 and M3 described in this paper.

That said, without a proper experimental design, like in your case, you can't really do anything

[SKY]

Thank you for the immense help. I am reading. I had no idea that such a wonderful resource is availabe.

[SKY]

In section 2.2.1

M4 models the ratio of a feature’s MTX (RNA) abundance to its MGX (DNA) abundance as a measure of ‘relative expression’ (i.e. transcripts detected per gene copy). These ratios are computed on smoothed, community-normalized MTX and MGX measurements and then log-transformed prior to model evaluation.

Are they telling to perform some other normalization prior to this method? I am not able to understand the difference between community and gene copy number normalization.

Please guide.

[SKY]

I think we usually say "Differentially Expressed" by comparing samples from one condition (say stress) with the other (say control). What if i have meta-transcriptome samples from only one condition e.g: water from a swimming pool X. How will i know which genes are Differentially Expressed by different microbes in the pool X?