can the same set of reads be sequenced in different experiments/runs?

[cantabile]

Hi all,

I'm a bit confused about the following data types on NCBI SRA site:
samples, experiments, runs

Specifically, I found from the dataset I download:
1. some samples are involved in more than one experiment
2. some samples are involved in more than 1 run with a single experiment

I've already browsed through this site
http://www.ncbi.nlm.nih.gov/books/NBK56913/,
but still not clear about some issues.

Suppose I have a sample X with 1000 reads:
1). if X is involved in experiments a and b, and it shows that a, b have different "libraries", can we assume anything about the reads sequenced in each experiment? (i.e. can both experiments sequence the same 1000 reads)

2). if X is related to runs y and z in experiment a, can we assume anything about the reads sequenced in each run? (i.e. can both runs sequence the same 1000 reads)

Any help would be appreciated!

[GenoMax]

You are "sampling" from the population (http://en.wikipedia.org/wiki/Sampling_%28statistics%29) to ultimately draw some statistical inference (http://en.wikipedia.org/wiki/Statistical_inference).

There is no guarantee that you are going to sequence the same "X" sequences from a library (if you run it again) but provided your library contents are "normally" distributed you should be able to compare the results from multiple runs.

If you are looking to get a certain "fold" coverage then you may end up running multiples lanes of sequencing with the same sample since you can only get a certain amount of sequence from each lane.

[cantabile]

Thanks for your reply!

My goal is to compare abundance counts between 2 different groups of samples.
I'm afraid that I might end up counting the same sequence twice if a sample is associated with more than 1 experiment or run.

So I guess it's ok to just use all experiments and runs for each sample?