Fig. 2From: MolData, a molecular benchmark for disease and target based machine learningMap of the bioassays’ descriptions using the output of the BioBERT model (The explained variance ratio for the first and second principal components are 14.72% and 5.40% respectively)Back to article page