Fig. 3: Benchmarking GENBAIT and other bait selection methods using NMF-derived metrics.
From: Computational design and evaluation of optimal bait sets for scalable proximity proteomics
Boxplots comparing GENBAIT, feature selection methods, and random selection based on mean scores: mean NMF Pearson correlation (a), mean NMF Cosine similarity (b), mean NMF KL divergence (c), NMF ARI (d), mean NMF GO Jaccard index (e); and min scores: min NMF Pearson correlation (f), min NMF Cosine similarity (g), max NMF KL divergence (h), min NMF purity score (i) and min NMF GO Jaccard index (j). Each method selected 30–80 baits over 10 random seeds. Box plots show median values; hinges are the 25th and 75th percentiles; whiskers indicate 1.5× the interquartile range (IQR). Highest average median values have been shaded in gray.
