### BFG Y2H Analysis Pipeline ###
**Requirements**
* Python 3.7
* Bowtie 2 and Bowtie2 build
### Files required ###
The pipeline requires reference files before running. They can be found on GALEN:
```
all reference files contain all the barcodes in fasta format
path: /home/rothlab/rli/02_dev/08_bfg_y2h/bfg_data/reference/
```
Before running the pipeline, you need to copy everything in these two folders to your designated directory.
#### Build new reference ###
If you need to build a new reference for your analysis, please follow:
1. You can refer to the create_fasta.py script to build the new fasta file
2. Make sure the name for the sequences follows the format: `>*;ORF-BC-ID;*;up/dn`. In other words, the ORF-ID should always
be the second item, and the up/dn identifier should always be the last item. (see examples below)
3. Example sequences in output fasta file:
```
>G1;YDL169C_BC-1;7;up
CCCTTAGAACCGAGAGTGTGGGTTAAATGGGTGAATTCAGGGATTCACTCCGTTCGTCACTCAATAA
>G1;YMR206W_BC-1;1.0;DB;up
CCATACGAGCACATTACGGGGCTTGAGTTATATAGTCGATCCGGGCTAACTCGCATACCTCTGATAAC
>G09;56346_BC-1;24126.0;DB;dn
TCGATAGGTGCGTGTGAAGGATGTTCCCCCGGTCACCGGGCCAGTCCTCAGTCGCTCAGTCAAG
```
4. After making the fasta file, build index with bowtie2-build
`bowtie2-build filename.fasta filename`
5. Update main.py to use the summary files you generated
* Edit parse_input_files() to add a case
### Running the pipeline ###
* Install from pypi (recommend): `python -m pip install BFG-Y2H`
* Install and build from github, the update.sh might need to be modified before you install
```
1. download the package from github
2. inside the root folder, run ./update.sh
```
1. Input arguments:
```
usage: bfg [-h] [--fastq FASTQ] [--output OUTPUT] --mode MODE [--alignment]
[--ref REF] [--cutOff CUTOFF]
BFG-Y2H
optional arguments:
-h, --help show this help message and exit
--fastq FASTQ Path to all fastq files you want to analyze
--output OUTPUT Output path for sam files
--mode MODE pick yeast or human or virus or hedgy or LAgag
--alignment turn on alignment
--ref REF path to all reference files
--cutOff CUTOFF assign cut off
```
2. All the input fastq files should have names following the format: y|hAD*DB*_GFP_(pre|med|high) (for human and yeast)
3. Run the pipeline on GALEN
```
# this will run the pipeline using slurm
# all the fastq files in the given folder will be processed
# run with alignment
bfg --fastq /path/to/fastq_files/ --output /path/to/output_dir/ --mode yeast/human/virus/hedgy --alignment --ref path/to/reference
# if alignment was finished, you want to only do read counts
bfg --fastq /path/to/fastq_files/ --output /path/to/output_dir/ --mode yeast/human/virus/hedgy --ref path/to/reference
```
### Output files ###
* After running the pipeline, one folder will be generated for each group pair (yAD*DB*)
* The folder called `GALEN_jobs` saves all the bash scripts submited to GALEN
* In the output folder for each group pair, we aligned R1 and R2 separately to the reference sequences for GFP_pre, GFP_med and GFP_high.
* `*_sorted.sam`: Raw sam files generated from bowtie2
* `*_noh.csv`: shrinked sam files, used for scoring
* `*_counts.csv`: barcode counts for uptags, dntags, and combined (up+dn)
