Manifest file for paired-end reads


I use demultiplexed paired-end libraries stored in multiple fastq.gz files.
Is there a way to add columns in manifest file to set experimental conditions (times, drug, sex) corresponding to each library ? Or i need to run qiime separately for each condition ?

thank you –


Hi! I would prepare a sample-metadata.csv file and include all the experimental conditions in it for use in q2 pipeline. Manifest file should be for importing all the .gz files into q2. You may like to read the tutorials at and

1 Like


is there a way to add metadata (condition, time, sex) when i import demultiplexed paired-end files ?
I have only demultiplexed files, so i use a manifest file, but manifest file is not dedicated to add metadata informations. So, how to specify metadata corresponding to each file ?

thank you.

@lmanchon — please see @bsen2018’s excellent advice. It sounds like you are conflating the sample metadata with the manifest file. These are separate files with separate purposes, see the links @bsen2018 posted to learn more.

yes i know but is there a solution to combine manifest file and metadata-sample files ?
Manifest file tells only where to get the input files but i need also to inform qiime with my metadata, don’t you ?
thank you –

No — they are always kept separate.

Correct. You use it for importing and then never again. The reason is that manifest files describe the location of individual sequences, and in the case of paired-end would have more lines than samples.

Correct — see the metadata tutorial that was linked above. You will create a separate file describing sample metadata.

Good luck!

ok, i have imported all my demultiplexed paired-end files using manifest-file, i have a big qza file (300Go).
And now which command i need to run to link my metadata file with the raw data ?

thank you –

Good! :champagne:

No need, that isn’t how metadata works in QIIME 2. Take a look at the Metadata tutorial for more info. Also, if you haven’t gone through the Overview, please start there!

1 Like

Ok, i understand, metadata file work only with multiplexed input files, and there is no way to use metadata file with demultiplexed paired-end input files.

Not quite — this statement is true for importing/demultiplexing, but sample metadata files can be used later on no matter what type of import you used.

Manifest files facilitate importing demultiplexed sequence files into a single QIIME 2 artifact by specifying where each sequence (and if paired-end also it’s pair) is located. So you import as a manifest format and there is no need to demultiplex.

Sample metadata files can contain all sorts of information about samples — including the barcodes used, in which case you would use the metadata file during demultiplexing to specify how a multiplexed sequence file (i.e., EMP format sequences) should be split. But we use sample metadata again in downstream analysis to store other information about these samples.

You said:
“…we use sample metadata again in downstream analysis to store other information about these samples.”

Yes, but how ? give me an example, i have a big demux-paired-end.qza file that contain all my data, which command do i use to integrate or merge my_metadata.tsv file with the qza file ?

read the tutorials, e.g., here.

None. At no point is metadata merged into your file. Metadata is kept separate, the purpose being that metadata should be flexible and easily modified.

It’s contradictory.
You tell me that it is not necessary to import it using a specific command.
But then, how does qiime2 know to use the metadata file?

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.