Tamipami: A web application and command line interface for finding the TAM and PAM sites of novel endonucleases
Carlos Orosco1, Pyush Jain1, Adam R. Rivers2
- University of Florida, Department of Chemical Engineering, Gainesville, FL USA
- United States Department of Agriculture, Agricultural Research Service, Genomics and Bioinformatics Research Unit, Gainesville, FL USA
When a new Cas or TnpB guided endonuclease is discovered or engineered, one of the first tasks is identifying its PAM or TAM recognition site. This can be done by treating a pool of plasmid DNA containing a target site adjacent to a random region. The random regions containing the PAM/TAM site are recognized and cut in the presence of the endonuclease and a guide RNA. These become depleted in the sequencing library. By comparing an uncut control library to a cut experimental library it is possible to identify the PAM/TAM site. The method was introduced by Walton et al. 2021. That work deposited the plasmid pools with Addgene, making the lab protocol accessible.
This web application and command line application builds on the work by creating software that simplifies the analysis of the sequencing data and adds rich interactive visualizations for selecting the PAM/TAM site.
The application can be installed into a conda environemnt using Pip and conda.
conda create -n tamipamienv -c conda-forge python=3.13
conda activate tamipamienv
conda install -c bioconda bbmap
git clone [email protected]:USDA-ARS-GBRU/tamipami.git
cd tamipami
pip install .
The web application can be found at https://tamipami.che.ufl.edu
To launch a the streamlit web application locally at http://localhost:8501 run this commmand:
streamlit run tamipami/app.py
| Tamipami Inputs | Tamipami results |
|---|---|
 |
 |
- Load the forward and reverse FASTQ files for a single control and experimental library in the four input boxes on the sidebar.
- Select the Addgene library used for the experiment. Or...
- If no library is selected, enter the target sequence and the orientation (5prime is PAM-Target, 3prime is Target-PAM (like spCas9)).
- Select the maximum length to analyze. You can compare all smaller lengths once you have analyzed the data
- Hit 'Submit'. This will process your files.
- After you have hit 'Submit' you can explore your data interactively. The key interface is the z-score slider bar. Moving that will set the cutoff value to separate kmers that cut from those that did not. This will update the histogram of sequence z-scores, the table of reads the degenerate sequences created and the sequence motif.
- Be sure to explore each length tab, to select the PAM/TAM site best supported by your data.
- Export the raw run data.
To facilitate automated analysis we have also released a command line version of TamiPami. The CLI has two subcommands: process and predict. The first step is to process the raw FASTQ data into counts of kmers. This data can be sent ot a JSON file or STDOUT. The second step is to predict the PAM Sequences given a specific cutoff value. The commands can be piped together.
TamiPami: a CLI application to parse High throughput PAM/TAM site sequencing data
positional arguments:
{process,predict}
process Sub-command "process": used to process FASTQ data into a summarized json output
predict Subcommand "predict" use to predict PAMs/TAMs and summary data for a selected length and cutoff value
options:
-h, --help show this help message and exit
--log LOG Log file
-V, --version show program's version number and exit
usage: tamipami process [-h] --cont1 CONT1 --cont2 CONT2 --exp1 EXP1 --exp2 EXP2 [--outfile OUTFILE] [--library {RTW554,RTW555,RTW572,RTW574}] [--spacer SPACER] [--orientation {3prime,5prime}]
[--length [3-8]]
options:
-h, --help show this help message and exit
--cont1, -c CONT1 A forward .fastq, .fq, .fastq.gz or .fq.gz file. .
--cont2, -c2 CONT2 A reverse .fastq, .fq, .fastq.gz or .fq.gz file.
--exp1, -e EXP1 A forward .fastq, .fq, .fastq.gz or .fq.gz file.
--exp2, -e2 EXP2 A reverse .fastq, .fq, .fastq.gz or .fq.gz file.
--outfile OUTFILE A path to a hdf5 file of the results, if missing binary data will be sent to STDOUT.
--library {RTW554,RTW555,RTW572,RTW574}
The Addgene library pool. For custom pools use the --spacer and --orientation flags
--spacer SPACER The spacer sequence for the guide RNA. Not needed if ---library is used
--orientation {3prime,5prime}
the side of the spacer the PAM/TAM is on
--length [3-8] The length of the PAM or TAM sequences
usage: tamipami predict [-h] [--input INPUT] [--cutoff CUTOFF] --predict_out PREDICT_OUT
options:
-h, --help show this help message and exit
--input INPUT An file containing the data from a TamiPami process run or downloaded data from the web ap, if not input is provided STDIN will
be assumed
--cutoff CUTOFF A json string containing the kmer lengths and the Zscore cutoff values above which kmers are considered part of the PAM/TAM.
Single and double quotes are required. If no cutoff is provided it will be automatically calculated using univariate k means
clustering. Example input: --cutoff '{"3": 2, "4": 2, "5": 2, "6": 2}'
--predict_out PREDICT_OUT
A file directory containing the PAM/TAM and degenerate sequences identified
Example datasets are available under NCBI Bioproject PRJNA1298332 : Tamipami: Software for determining the PAM and TAM sites of new CRISPR/Cas and TnpB nucleases . For more information see data/readme.md.
Detailed performance metrics for 7 datasets are found at benchmark/readme.md. TL;DR most datasets take 15-30 seconds to process.
If you need help or encounter errors please request support on the Tamipami Github issues page
Docker deployment command:
docker run -d \
-p 8501:8501 \
--name tamipami-app \
streamlit-app:latest \
streamlit run /app/tamipami/app.py --server.port=8501 --server.address=0.0.0.0
This software is a work of the United States Department of Agriculture, Agricultural Research Service and is not copyrightable under U.S. Code Title 17, Section 105. It is released under a Creative Commons CC0 public domain attribution.