Skip to content

Latest commit

 

History

History
53 lines (38 loc) · 2.84 KB

Tutorial.md

File metadata and controls

53 lines (38 loc) · 2.84 KB
Raw

Getting Data

We have parsed data collected from American Gut Project and placed the data, which has been split up by sample site, in a new Git repository. We shall use the data at Gregory Ditzler's data repo. To get the data, run:

git clone https://github.com/gditzler/DataCollections.git
cd DataCollections/
DCPATH=$(pwd)

Running an Experiment with Fizzy and NPFS

Once the data is downloaded, edit run-fizzy-experiments.sh to fill in the path variable for DCPATH in the script. This script provides a fairly comprehensive collection of experiments on how fizzy and npfs can be used. All result files will be saved in outputs/.

Running Fizzy on a subsample of Omnivores and all Vegans

This experiments is a bit more involved than some of the previous ones because we are required to modify the map and Biom files. First we should split the data by their diet. To do this, we use QIIME's split_otu_table.py.

nsel=50
cd ${DCPATH}/AmericanGut/
 
# split the otu table in to different diet types
split_otu_table.py -i AmericanGut-Gut-Diet.biom -m AmericanGut-Gut-Diet.txt -f DIET_TYPE -o tmp/
cd tmp/

Once the split performed, we need to 'subsample' the Omnivores data, such that we have 65 samples. To do so, lets modify the map file such that the first 65 entries that their original diet, but the others are going to be marked as being dropped. The AmericanGut-Gut-Diet_Omnivore.biom and its new map file are going to be split again.

# create a new omnivore mapping file to identify a subsample of size 65 
cat -n mapping_Omnivore.txt | awk -F '\t' '{if($1 < 66){ print $2,$3;}else{print $2,"DROP"} }' | tr ' ' '\t' >  new_mapping_Omnivore.txt
split_otu_table.py -i AmericanGut-Gut-Diet_Omnivore.biom -m mapping_Omnivore.txt -f DIET_TYPE -o subset-omni/

Now we need to merge the subsampled omnivores data with the vegans, so we are going to merge the OTU and map files.

# merge the otu table of the subsample of omnivores and vegans
merge_otu_tables.py -i subset-omni/AmericanGut-Gut-Diet_Omnivore_Omnivore.biom,AmericanGut-Gut-Diet_Vegan.biom -o subsample-OmniVegan.biom
 
# merge the mapping files of the subsample of omnivores and vegans
merge_mapping_files.py -m subset-omni/mapping_Omnivore.txt,mapping_Vegan.txt -o subsample-OmniVegan.txt

Now that this is completed, we can run fizzy

# run fizzy 
fizzy -l DIET_TYPE -f JMI -i subsample-OmniVegan.biom -m subsample-OmniVegan.txt -n ${nsel} -o subsample-OmniVegan-Fizzy-JMI.txt -r subsample-OmniVegan-Fizzy-JMI.biom
 
# print out the differences in the abundances
cat -n subsample-OmniVegan-Fizzy-JMI.txt | tr -d ' ' | awk -F '\t' '{d=$4-$7; if (d >= 0) {print $1-1,$2,sprintf("%.9f", d),"p";}else{print $1-1,$2,sprintf("%.9f", -d),"n";} }'  | tail -n +2 | sort -k3 -nr | head -15