Is it possible to use isOnclust on multiple samples?

[DesmondoDekker]

Hi! I am working with samples from nanopore. I successfully used isOnclust on a single sample, but I would need to compare various samples. Is it possible to use isOnclust on multiple samples?

[ksahlin]

It depends on what your input data is and what is the desired output of the analysis.

Are all the reads within one file but with sample-specific barcods? maybe then perform an initial barcode clustering/trimming with e.g. pychopper, and the run isONclust on each of the sample specific files.

Otherwise please describe the data and the desired outcome in more detail.

[DesmondoDekker]

Hi, thank you very much for your prompt answer. I have 10 different singles file.fastq. I would like to use isOnclust, but then I would like to compare one sample against another one. The point is that if i run isOnclust on every single sample, I do not know how to compare the clusters across the samples. Brefly, I would like only one "otu table" with 10 samples. I hope I have been clear.

Thanks a lot

[ksahlin]

Not sure it is possible to create an OTU table without some scripting. One idea:

1. Label all your reads in the headers (sample 1 gets a _1 appended, etc) for uniqueness.
2. Combine all reads into one big file.
3. Cluster the big file with isONclust.
4. Parse the isonclust csv output file. The file contains a line for each read and which 'cluster representative' it is part of. Summing the reads per sample from this file, it will give you the abundance of each sample in a cluster.

[DesmondoDekker]

It sounds good. Thanks for the suggestion!