Use Case: Disease Classification

Challenge: What Kind of Infection?

  • Many medical offices have few testing options for discovering disease source.
  • Providers can be unsure if an infection is viral or bacterial.
  • Lack of clarity leads to improper prescription of antibiotics.

Solution: Look for the Body’s Response

  • Bacterial, viral, and fungal infections each cause different genetic responses in the body.
  • Use blood test to measure this genetic response and find the infection type.
  • Use supervised learning to model the probabilities of viral infections, bacterial infections, etc.

Impact: A Better Diagnosis

  • New insight about previously untested medical conditions.
  • Gives patients and doctors better clarity about care.
  • Reduces inappropriate antibiotic use, saving money and improving public health.

Infinia ML can help you find insights in data.
Email us to learn more.

The relative expression of gene 1 is the most important factor in classifying whether the patient has a bacterial, viral, or fungal infection, has a non-infection immune response, or is healthy.

The distribution accounts for variance across multiple rounds of model training.