Frequently Asked Questions
What is VaccinesAI?
VaccinesAI is an open-source web application with two reverse vaccinology pipelines: VacSol-ML(ESKAPE), targeting ESKAPE pathogens, and Bvac-AI, addressing a broader range of antimicrobial-resistant pathogens. It uses machine learning to identify vaccine immunogens.
What are ESKAPE pathogens?
ESKAPE pathogens are six bacteria (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, Enterobacter species) known for antimicrobial resistance.
How does Bvac-AI differ from VacSol-ML(ESKAPE)?
Bvac-AI is an enhanced version of VacSol-ML(ESKAPE), targeting a wider range of antimicrobial-resistant pathogens beyond ESKAPE bacteria, with the same machine learning approach.
What type of input does the application require?
The application requires protein sequences in FASTA format, uploaded as a .fasta file.
What analyses are performed on the input sequences?
Analyses include MHC-I and MHC-II binding (netMHCpan, netMHCIIpan), B-cell epitope prediction (Bepipred), physicochemical properties, signal peptide prediction (SignalP), and machine learning predictions.
What is reverse vaccinology?
Reverse vaccinology uses computational analysis of genomic/proteomic data to identify vaccine candidates, automated by VaccinesAI with machine learning.
What is the output of the application?
Outputs include a dataset.csv file with predictions, probabilities, and features, plus a web interface with confidence messages.
What does the confidence message mean?
Confidence messages indicate vaccine candidacy likelihood: Negative (0–50%), Low (50–60%), Moderate (60–75%), High (75–90%), or Very high (90–100%).
Can I use the application for non-bacterial pathogens?
Bvac-AI is optimized for bacterial pathogens, so results for non-bacterial sequences may be less reliable.
What tools does the application use?
Tools include netMHCpan, netMHCIIpan, Bepipred, SignalP, ProteinAnalysis, Peptides, and a stacking classifier.
Frequently Asked Questions (Continued)
Is the application free to use?
Yes, VaccinesAI is open-source and free, though usage quotas may apply on some platforms.
How can I interpret the MHC binding results?
MHC results include EL_Rank (lower is better), Strong Count (peptides with EL_Rank ≤ 0.5 for MHC-I, ≤ 2.0 for MHC-II), and Strong Fraction (proportion of strong binders).
What are signal peptides, and why are they important?
Signal peptides direct proteins to cellular compartments, indicating surface exposure, a key trait for vaccine candidates.
How long does the analysis take?
Analysis takes a few minutes for small datasets, with 16 steps tracked via progress updates.
Can I download the results?
Yes, results are saved as dataset.csv and downloadable via the web interface.
What should I do if I encounter an error?
Check the error message, ensure correct FASTA format, and contact support for persistent issues.
Does the application support batch processing?
Yes, multiple sequences in a FASTA file are analyzed concurrently.
How accurate are the vaccine candidate predictions?
Accuracy depends on training data and input quality, with confidence levels guiding validation.