JADBio is automated machine learning built for biological and molecular data, aimed at biomarker discovery. It offers both a GUI and an API, and is available through the AWS Marketplace.
Specialising matters here in a way it does not for general AutoML. Omics data has properties that break standard tooling: thousands of features against tens of samples, strong correlation between features, and a very real risk of finding a signal that is an artefact of the experiment rather than the biology. Tooling that does not know it is looking at molecular data will happily report an overfit result with a beautiful accuracy score.
It also covers survival analysis, which is standard in clinical research and absent from most general platforms. The obvious caution is one the field already knows: a biomarker discovered computationally is a hypothesis, and validation in an independent cohort is what makes it a finding.
