Re: Large study
Reply #1 –
I will try to answer to the best of my knowledge.
This number of samples are in themselves not a problem. But your files are rather large. This means that the peaklist might get too big if a large number of features can be detected. You can increase the intensity threshold to get around this. But then you don't get much for the sensitivity of your instrument that you paid for. I think the R problem of large vectors was solved with R v3. So the problem you will have is the relatively low amount of memory you have. Memory is quite cheap so you should probably investigate how much memory can be put in the computer you have.
You could try to analyse a few samples, observed the amount of memory used, and extrapolate from that to your sample size (use real samples, not blanks or QCs with fewer compounds) to see how much memory you need. Currently using more than one core eats a disproportionate amount of memory so keep that in mind.