QIIME2 pipeline

In order to extract and wrangle sequence data, please follow the following procedures.

1. Clone QIIME2 pipeline

git clone https://github.com/TracyRage/qiime_pipeline.git

Make all the .sh files executable

chmod +x *.sh 

2. Extract barcodes from raw sequences

In order to proceed with QIIME2 analysis, you need to create a metadata tsv file with barcodes attributed to each sample. Given the fact, that there’s already a metadata.tsv example in this repo, feel free to modify it, and add there your sample abbreviations and barcodes.

• Create a temporary directory in seqs/ and decompress raw *fastq.gz files

cd seqs/ && mkdir temporary_dir && gunzip *gz && cp *fastq temporary_dir && gzip *fastq && cd temporary_dir

• Install USEARCH

• Extract barcodes with USEARCH

# Do it for each sample
touch labels.txt && usearch -fastx_getlabels YOUR_SAMPLE_NAME.fastq -output labels.txt | head -n 3 labels.txt

In order to avoid redundance, extract barcodes from FORWARD sequences (i.e. R1)

• Change sample names and add barcodes in metadata.tsv. Optionally, change the other metadata entries.

• Don’t forget to properly rename your sequence files. Otherwise, QIIME2 won’t see them. Please, see example files in seqs/.

• Rename sample, barcode and forward (R1) / reverse (R2) fields: sample_barcode_L001_{R1 or R2}_001.fastq.gz

• If you done, please, delete example files.

3. Import data in QIIME2

bash import.sh

4. Process data

Go to QIIME View and check your demuz.qzv file, and decide what portion of sequnce to trim (median >= 28). Usually, default settings in process.sh are good enough. So, you may ignore this step.

bash process.sh

5. Train sequence model

This pipeline uses GTDB database.

If you have any other primers to work with, open training.sh and modify FORWARD and REVERSE variables. If not, just run the script.

bash training.sh

Analyze your dataset

Go to QIIME View and check your feature_table.qzv file, write down median frequency and feature count of the sample with the fewest count number. Please open analyze.sh and modify SAMPLING_DEPTH and MEDIAN variables.

bash analyze.sh

Conclusion

To see the bacterial distrbution, go to QIIME View and check your tax_bar_plots.qzv file.

For further statistics / graphics generation consult sping_analysis.Rmd (which is optional).