Extended Data Fig. 1: Integrative analysis of multiple TF SELEX datasets produces consensus binding models. | Nature Biotechnology

Extended Data Fig. 1: Integrative analysis of multiple TF SELEX datasets produces consensus binding models.

From: Prediction of protein–ligand binding affinity from sequencing data with interpretable machine learning

Extended Data Fig. 1: Integrative analysis of multiple TF SELEX datasets produces consensus binding models.

(a) Schematic contrasting ProBound’s multi-experiment learning strategy that builds a consensus model for a TF by simultaneously training on all relevant SELEX data for the TF with the traditional approach that builds independent models for every individual dataset. (b) Generalization performance of consensus binding models (y-axis) and single-experiment models (x-axis) on three different metrics (scatterplots). Points correspond to models trained on individual experiments and lines connect experiments used to build the corresponding consensus model. Points above the diagonal correspond to instances where the consensus model outperforms single-experiment models.

Back to article page