ProteinTailor

Species-specific codon optimization for increased heterologous protein expression.

Installation

To install the project, follow these steps:

1. Clone the repository: git clone https://github.com/ndreey/ProteinTailor.git
2. Install the dependencies manually or use: pip install -r requirements.txt (Note, requirments.txt must be located in the working directory)

MAC and Linux users, be aware that ProteinTailor will run into issues loading the ProteinTailor Report automatically as the pathfinder function was developed for a Windows OS. Successful installation and run are confirmed in Windows 11.

User Manual

After main.py has been executed, the ProteinTailor GUI will appear. Input format can be selected with the drop-down menu.

• cDNA
• mRNA
• amino acid sequence (aa-seq)
• Uniprot accession

ProteinTailor can process inputs in either FASTA format or as a pure sequence, though it is important to note that this version is limited to handling a single input at a time. The NCBI taxonomic identifiers (taxid) for the organism from which the protein originates and the E. coli strain must be entered to run ProteinTailor. It is recommended to provide a unique job title for each tailor run, as the ProteinTailor Report will be automatically opened in a new web browser tab. Each run overwrites the previous files in the temp/ directory (.HTML, .png) but is still accessible as long as the web browser tab is open. The HTML report and plots can be saved from the tab by right-clicking and pressing "Save as" or "Save image as". This allows for a streamlined approach where multiple reports can be generated each session.

The report will state the Job Title and provide the Tailored Sequence. In the Mirror Check, the raw sequence and tailored sequence are compared (NOT ALIGNED). The Statistics Table will hold the Codon Adaptation Index (CAI), GC-content, sequence length, number of unviable codons, and number of codons that were tailored. Further, two plots will be displayed showing the GC-distribution (across 80bp) of the raw and tailored sequence.

ProteinTailor Report Example

How it works

ProteinTailor tailors the sequence through four steps.

1. Input: If a Uniprot accession is provided, the amino acid sequence is retrieved using the Uniprot API. Regardless of the input type, the input sequences are converted to mRNA. The python_codon_tables package is used to store the codon usage tables (CUTs) from the CUT database kazusa.or.jp for each taxid.
2. Codon Tailor: The sequence is codon optimized by selecting the most frequent codons from the host CUT. If the organism's codon does not exist in the host CUT, the aa is determined using the organism CUT. Then by mapping the aa to the host CUT, the most frequent codon is chosen.
3. Final Fitting: The sequence is checked for false initiations and nonsense mutations introduced by the Codon Tailor step. This is done by locating the rogue start and stop codons. False initiations are removed by disrupting Shine-Dalgarno sites that are 6-12bp downstream of start codons. Stop codons in the reading frame are defined as nonsense mutations and will be stripped from the sequence. Frameshifted stop codons are ignored.
4. Protein Pickup: Presents the tailored sequence, "mirror check", statistics table, and plots using an HTML template and opens a web browser tab to make it possible to interact with the report.

Version 0.1.0

The current version includes a functional GUI but has only been tested for use with E. coli or similar prokaryotes as the host on a Windows operating system. It is currently capable of utilizing the algorithm parameters of codon bias, false initiations, and nonsense mutations. Please note that ProteinTailor will be continually updated as the algorithm is developed and its performance is evaluated with different organisms, including eukaryotes.

Troubleshooting

When running a ProteinTailor optimization there is a live terminal log to help you see where it gets stuck. As mentioned, MAC and Linux is inclined to form issues with opening and even creating the report. Importantly, the nucleic sequences can only be comprised of CODONS. This is the most common issue when running mRNA or cDNA sequences. Run this sequence, X70508, to see the error message when sequence is not comprised of codons.

To test that ProteinTailor works, test run it with these inputs.

• UniProt accession: P01308 or cDNA sequence: BC005255
• Homo sapiens taxid: 9606
• E. coli K-12 taxid: 83333

File Manifest of protein_tailor/

• main.py: the main program file
• gui/: a directory containing the code for the graphical user interface (GUI)
  • init.py: An empty file used to mark the gui directory as a Python package.
  • gui.py: The code for the GUI.
• program/: a directory containing the protein_tailor.py module which orchestrates the protein tailor program.
  • init.py: An empty file used to mark the gui directory as a Python package
  • protein_tailor.py: Code for the ProteinTailor class that utilizes the tailor_tools.
• tailor_tools/: A directory containing the code for the tailor tools used by ProteinTailor.
  • init.py: An empty file used to mark the tailor_tools directory as a Python package.
  • codon_tailor_tools.py: Code for the codon tailor tools.
  • final_fit_tools.py: Code for the final fit tools.
  • input_tools.py: Code for the input tools.
  • report_tools.py: Code for the report tools.
  • sequence_tools.py: Code for the sequence tools.
  • statistics_tools.py: Code for the statistics tools.
• temp/: A directory to store temporary files.
  • mirror.txt: Visual comparison between raw and tailored sequence.
  • ProteinTailor_Report.html: The HTML file presenting the ProteinTailor report.
  • raw_gc.png: Plot of the GC distribution across 150 bases of raw sequence.
  • tailored_gc.png: Plot of the GC distribution across 150 bases of tailored sequence.
• resources/: A directory containing resources used by the program.
  • protein_tailor_bg.png: Background image for the ProteinTailor HTML report.
  • protein_tailor_icon.png: Icon image for the GUI.