Vaginal 16S foetal alcohol spectrum disorder

# Wiki

# Data location

This project was run on the Ilifu server

```c

/cbio/users/katie/Hemmings

```

## Testing data raw reads

```c

/cbio/users/katie/Hemmings/fastqs

```

## EMIRGE software setup

[[https://github.com/csmiller/EMIRGE]] software was setup as a singularity container, using a Docker image available on Dockerhub [[https://hub.docker.com/r/golob/emirge]].

1. From CBIO's BST server (katie@bst.cbio.uct.ac.za) I pulled the docker container with ```docker pull golob/emirge```

2. The singularity container was built with ```rsync -avvP -e "ssh -i /home/katie/.ssh/id_rsa" singularity-containers/emirge_latest.simg katiel@transfer.ilifu.ac.za:/cbio/users/katie/Hemmings/containers```

3. The singularity container was transferred to Ilifu with  ```rsync -avvP -e "ssh -i /home/katie/.ssh/id_rsa" singularity-containers/emirge_latest.simg katiel@transfer.ilifu.ac.za:/cbio/users/katie/Hemmings/containers```

## EMIRGE troubleshooting

Unfortunately there were several hurdles setting up the SSU database for use with EMIRGE. Firstly, the ```emirge_makedb.py``` in the singularity container did not work. I had to do a git clone on the original github repo (https://github.com/jgolob/EMIRGE) and edit the FTP site specified in the script from SILVA_{rel}_SSURef_Nr99_tax_silva_trunc.fasta.gz to SILVA_{rel}_SSURef_NR99_tax_silva_trunc.fasta.gz (Nr99 to NR99). This script is available at ```/cbio/users/katie/Hemmings/EMIRGE_master_branch/EMIRGE/emirge_makedb.py```

```./emirge_makedb.py -p8 --silva-license-accepted```. The steps executed in ```emirge_makedb.py``` are:

1) download the most recent SILVA SSU database, 2) filter it by sequence

length, 3) cluster at 97% sequence identity, 4) replace ambiguous bases

with random characters and 5) create a bowtie index.

Still, the above command aborted after the clustering step (step 3), with error:

```Traceback (most recent call last):

  File "./EMIRGE_master_branch/EMIRGE/emirge_makedb.py", line 438, in <module>

    main()

  File "./EMIRGE_master_branch/EMIRGE/emirge_makedb.py", line 425, in main

    randomized_fasta = randomize_ambiguous_fasta(clustered_fasta,

  File "./EMIRGE_master_branch/EMIRGE/emirge_makedb.py", line 339, in randomize_ambiguous_fasta

    outf.write(randomize_ambiguous(line.rstrip("\n")))

TypeError: a bytes-like object is required, not 'str'```

I therefore had to manually replace ambiguous characters in the resulting clustered DB named ```SILVA_138.1_SSURef_NR99_tax_silva_trunc.ge1200bp.le2000bp.0.97.fasta```, using the script from the github site (this script was not available in the singularity container, but can be found under ```/cbio/users/katie/Hemmings/EMIRGE_master_branch/EMIRGE/utils/fix_nonstandard_chars.py```. The executed command was ```python2 ./EMIRGE_master_branch/EMIRGE/utils/fix_nonstandard_chars.py < SILVA_138.1_SSURef_NR99_tax_silva_trunc.ge1200bp.le2000bp.0.97.fasta > SILVA_138.1_SSURef_NR99_tax_silva_trunc.ge1200bp.le2000bp.0.97.fixed.fasta```. Next, a bowtie index was built for this fasta file with ``` bowtie-build SSU_candidate_db.fasta SSU_candidate_db_btindex``` where the bt index was named SILVA_138.1_SSURef_NR99_tax_silva_trunc.ge1200bp.le2000bp.0.97.fixed.bowtie*. Note that these commands were still executed from within the Singularity container to make use of the necessary software installed therein.