I thought about this question when did my sampliing work, however, each sample collected are all by random in my view, whether mixed scitifically or not, thus the results of the sequencing, though hard to concede, are only reflections of the ever-present microbial structures.

Sidan,

Thank you very much for your answer. Yes, it's really important that the samples are collected randomly. About if the samples could be mixed, I think should consider the below conditions.

1. Sampling scale. If your study are local scale and with few sampling sites, I think it's better not to mix the sub-samples. If your study are regional or global scale and with large number of sampling sites, I think you can mix the sub-samples.

2. Sample heterogeneity. Of course, if the sub-samples are large heterogeneous in a sampling site, you should not mix them.

In my view, if the sub-samples could be mixed depending on your study aim, the sampling scale, number and heterogeneity.

Thank you again.

The below link has discussion related to this question:

http://mb3is.megx.net/gustame/warnings/warning-pseudoreplication

Hi Guangshan,

       Thank you for your reply and more thankful for the link you provided.

       Let's keep in touch.

Hi Wei

Your sampling should reflect the research question you want to ask,  As you know, soil is full of microsites, each with is own microbial diversity. So depending on what you wish to know you have to sample accordingly. It is alo necessary to validate your experimental methods to ensure that they are giving you a true representation.

Best wishes

Chris