Fig. 1: Comparison between PROXIMAL and PROXIMAL2.
From: Extending PROXIMAL to predict degradation pathways of phenolic compounds in the human gut microbiota
a Percentage accuracy in recovering annotated products in KEGG for PROXIMAL and PROXIMAL2 in three different scenarios: Case1, Case2 and Case3. In Case1 and Case2, we consider reactions with a single reaction centre. We do not consider atom charge information in Case1, but we do in Case2. In Case3, we consider all the reactions in the KEGG version used in PROXIMAL. The number of reactions used in each of the cases is shown in parenthesis. b Heatmap representing the chemical similarity of the predicted and annotated products in KEGG in Case 1 and 2 where at least one of the two algorithms fail to predict the annotated product. c Barplot representing the reaction coverage with look-up tables and those correctly predicting the annotated product in KEGG reactions. The y- axis shows the coverage in percentage. The total number of reactions are indicated over the bars.
