GC content in metagenomic data

Question

[bioinfo]

I have sequenced a few metagenomes on Illumina (DNA extracted from fish tissue). The average number of reads I got is around 40 million/sample. I have done the preliminary QC on the datasets, and I can say that the quality of the dataset is quite high (based on the FastQC report) but I see a clear GC bias (only 35-37% GC contents). Can anyone guide me whether it is due to species composition of the samples or PCR amplification during library prep or later in the cluster generation step. Is there any other reason for this GC bias in metagenomic samples?

[Joe]

Do you know what the expected GC of your likely organisms are? GC content is rarely if ever 50:50, so that doesn't sound super out of the ordinary so far...

[Carambakaracho]

What sort of metagenomic samples are you extracting from tissue? Metagenomic samples are a more or less complex and diverse collection of organisms, some of which might be low GC.