I´m new to qiime and I don’t understand anything on this forum. Forgive me if I sound ignorant but I really need help and my supervisor is asking me to check on the forum on how to combine 3 sequence runs as one. I just want to ask for advice since we run 3 chips already. (1st chip has few good sequence count, 2nd chip has also few sequence count and the 3rd chip has a 75% readable sequence count.)
So what I did was to remove or eliminate the bad sequence count (anything below 1,700 sequence count) by editing the metadata file and choosing the samples which are the good one and re-running them again. But my main problem is how do I combine all the 3 chips ( which have the good sequence count) into 1? How do I merge them??
If someone could explain in a comprehensible way then that would be perfect. And also pls. let me know on what code to write on qiime2 to merge all the three. Thank you so much for your help!
Follow along with this tutorial, specifically the "merging" step that I am linking to below. This tutorial gives an example of how to merge multiple sequencing runs after denoising. If you are importing the raw fastq data into QIIME 2, you can follow all of the steps leading up to merging in this tutorial (whereby each sequencing run is denoised independently). If you have already done denoising outside of QIIME 2, import the results of each run into QIIME 2 as a FeatureTable[Frequency] and then use the merge method to merge these features tables into one: https://docs.qiime2.org/2019.4/tutorials/fmt/#merging-denoised-data
There are lots of great tutorials on the QIIME 2 website, including some intended to orient beginners. Read through this one, which gives more of a conceptual overview, and then proceed to the others in the tutorial to get a better understanding of how to use QIIME 2 powerfully: https://docs.qiime2.org/2019.4/tutorials/overview/
Yes, you should, if only because this is a good opportunity to make sure everything is correct and you have all the same metadata across runs. Make a single TSV out of the 3 and your life will be much easier after that!
In general, no: the barcodes do not matter at this point, since you have already demultiplexed.
HOWEVER:
Yes! Change the sample IDs so that all samples have unique IDs. In general, integers make bad sample IDs — use a more readily identifiable or at least unique sample ID. Using an integer ID will not only result in naming conflicts like this, it can also cause some downstream analyses to bug out (and I don't mean just in QIIME 2 — I mean everywhere: if you use R, statistical software, even Excel for downstream analyses)