20230605
- Fixed the inability to recognize lowercase characters in uploaded fasta files.
- Please upload sequence no larger than 20,000,000 in length.
20240904
- For security reasons, the web server has recently been migrated.
- All models are trained on whole genome sequences. Currently, the minimum length of the web server input sequence is 1,000,000.
Recent and ongoing developments in microbiome science and Machine learning are enabling new frontiers of research for probiotics and prebiotics.
Understanding strain-specific of probiotic offers a striking opportunity for carry out further studies in the probiotic mechanism of probiotics.
ProbAI aims to provide an online probiotics predictors tool.
ProbAI contains three support vector machine(SVM) models based on different data sets(Probiotic Predictor,Lactobacillus, Bifidobacterium and other classifiers,Probiotic Lactobacillus Predictor).
Users can choose different models to input the microbial genome sequence (fasta), and obtain the probability that the microbe is a probiotic, the probability that it is a probiotic lactobacillus, a probiotic bifidobacterium, the probability of other probiotics, and the probability of a probiotic lactobacillus.
And we also explore gene expression patterns of the probiotics dataset we train the Probiotic Predictor model.
ProbAI contain more than 2282 probiotics strain enriched genes.