Prediction
When should you use prediction?
Metagenomic sequencing is a powerful tool to obtain data from a sample, yet the amount of data it produces is enormous and far from clear-cut compared to (some applications of) PCR, where a difference in color can already answer the question you are working to answer.
With the help of artificial intelligence, we can provide you with a similar clarity, while using all of the provided data. On top of a detailed taxonomy and functional overview of your sample, we can include a prediction of the sample's phenotypes, tuned to your specific needs.
For example, if you are interested in the presence of a specific pathogen, we can provide you with a prediction of the presence of that pathogen in your sample. If you are interested in identifying the presence of a specific metabolite, we can provide you with a prediction of the amount of that metabolite in your sample. If you are interested in the presence of a specific antibiotic resistance gene, we can provide you with a prediction of the presence of that gene in your sample.
Caveats
When working with Artificial Intelligence, there are a lot of ways to improve your results. At Hologenomix, we work constantly to help improve these factors.
Quality and quantity of samples matter. In general, more data will render AI better able to generalize towards unseen cases. We collect data from a wide range of domains to make generalization easier across many different domains.
Advanced models require a lot of tuning. Simply running a model on your data will not give you the best results. We constantly work on tuning our models to increase the accuracy of our predictions.
Prediction results can be hard to interpret. With a wide range of metrics from accuracy to Kolomogorov Smirnov charts, understanding what the odds are of the predictors being correct can be difficult yet crucial. At HologenomixTM, we make this interpretation easy with handy visualizations. Please contact us for an example!
Interested? Please contact us!