dhimmel / fratjuice

Uncovering the microbes of fraternity basements

The #fratjuice microbiome

For study background, see the inaugural tweet.

Scripts

generate-details.py creates markdown documents with details for each sample.

Environment

This repository uses conda to manage its environment as specified in environment.yml. Install the environment with:

conda env create --file=environment.yml

Then use source activate fratjuice and source deactivate to activate or deactivate the environment. On windows, use activate fratjuice and deactivate instead.

fratjuice's Issues

Registering samples on the uBiome Explorer

I messaged Elies regarding the status of our samples. It turns out, according to Elies, uBiome processed the samples on 2017-10-11! She wrote:

Let me know if you can access the data from the Explorer website or not. There was no user associated with the samples in the database, so maybe you never registered them?

It didn't occur to me to register these kits in the Explorer, which is why I didn't notice the processing was complete. So I just registered the samples in the Explorer now on 2017-12-13.

Methods

I clicked on "Activate Kit" and entered the kit info from kits.tsv. For each sample (kit-tube pair), I entered the date as 2017-07-20 (for when @bemert transferred the samples into the uBiome tubes, see #2 (comment)). Then I changed the sample type to custom from gut/spare. Custom is documented as:

Custom samples are a special category for non-standard sampling, e.g. sampling your pet. Is this a custom sample?

In the notes section, I added the sample information. This was a manual process, so I was careful not to make errors. Elies sent me some data, which we can use in the future to confirm that our sample-to-tube assignments are in concordance.

Here is a screenshot of the 10 samples on the uBiome Explorer:

Note that we have data on both the concentrated and unconcentrated preparations, which hopefully will provide some comforting redundancy.

Preparing the fratjuice for uBiome fratseq

uBiome — the 23andme of metagenomics / the microbiome — has graciously agreed to sponsor the 16S ribosomal RNA sequencing of our initial 5 fratjuice samples.

This issue will discuss our sample preparations for the uBiome kits.

Extracting species abundance from uBiome FASTQ outputs

We now have sequencing data from uBiome as both raw FASTQ sequences and taxonomy-level abundance summaries. We'd like to compute taxonomical abundance from the FASTQ data using an open source pipeline, as part of this repository. I haven't done anything like this before.

@eliesbik mentioned that the programs mothur (mothur/mothur) or Quime could be useful. Here's a comparison blog post. Note that Qiime version 1 (qiime/qiime) will soon be replaced by Qiime2 (qiime2/qiime2). @eliesbik also mentioned that someone familiar with the process at uBiome may be interested in contributing.

There's a ubiome-opensource/microbiome-tools repository by @richardsprague, which contains information on analyzing uBiome data. However, it looks like this project currently focuses on analyzing the taxonomic summaries rather than the raw FASTQ data.

On Biostars, there was a question about processing uBiome FASTQ data, but the answers are inconclusive.

Anyways, we'll use this discussion for coordinating how to process the FASTQ data.

Experiments to run on the initial samples

Currently, we have five samples. Each sample is in either a 15 or 50 mL conical-bottom tube. Some samples have sediment (i.e. dirt) which has settled to bottom. Samples are currently under refrigeration, with a target temperature of 4 degrees Celsius as per the recommendations of the Deb Hogan Lab.

So here are some experiments that could be interesting:

• alcohol concentration: much of the fratjuice consists of Keystone Light (which is ~4.1% abv).
• 16S rRNA sequencing to quantify prokaryotic microbes
• 18S rRNA sequences to quantify small eukaryotes
• Plating to see the cultures that grow

All ideas are welcome.

Best methodes for plating fratjuice

We recently froze some of the samples in preparation for 16S fratseq: see #2 (comment). However, we still have some refrigerated samples that we were thinking of plating. One goal would be to generate some nice Petri dish photos that help communicate the experimental design.

Does anyone have advice on what types of plate media we should use? Is a basic agar plate a good start?

Accessing fastq files from raw uBiome zips

I am having trouble accessing the fastq files that are in the raw uBiome zip files. Using standard file extraction doesn't work because unzipping XXXXX.zip will produce XXXXX.zip.cpgz. Attempting to open XXXXX.zip.cpgz produces another XXXXX.zip file, leading to a frustrating cycle.

In the "fastq" branch of the repository, there are fastq.gz files for forward and reverse reads that I was able to open and analyze in R, so there must be a way to access these files--I just cannot figure it out.

fratjuice deepart: using deep learning to combine frat and bacteria pics

I've been using https://deepart.io to combine our fraternity/fratjuice pics with our petri dish pictures.

I've done 3 tries thus far: UDbF1AYs, JHHdlUxf, FhPwHeGU. For example, the following art melds the style of frat to the photo of the petri dish.

@bemert I think we need petri dish pics with more color variation and contract. Do you think we could get this with microscopy / staining? How do we beautify?