RAVEN 2.0: a versatile toolbox for metabolic network reconstruction and a case study on Streptomyces coelicolor

Supplementary Information to the paper "RAVEN 2.0: a versatile platform for metabolic network reconstruction and a case study on <i>Streptomyces coelicolor</i>", as posted on bioRxiv.<br>RAVEN 2.0 is available from GitHub.<br>Sco4 is available from GitHub.<br><b>S1 Data. Model structure after merging of the three <i>de novo</i> draft reconstructions. </b>The MetaCyc- and KEGG-based <i>de novo </i>reconstructions for <i>S. coelicolor </i>are merged using the <i>combineMetaCycKEGGModels</i> function, as detailed in the Sco4Reconstruction script available from the Sco4 GitHub repository under ComplementaryScripts.<br> <b>S1 Table. Mapping of reactions from iMK1208 to MetaCyc and KEGG identifiers.</b> To facilitate comparison of the existing iMK1208 model and the <i>de novo </i>reconstructions, iMK1208 reactions were mapped to their respective KEGG and MetaCyc identifiers. This table is used in the Sco4Reconstruction script available from the Sco4 GitHub repository under ComplementaryScripts.<br> <b>S2 Table. Mapping of metabolites from iMK1208 to KEGG and MetaCyc identifiers. </b>To facilitate comparison of the existing iMK1208 model and the <i>de novo </i>reconstructions, iMK1208 metabolites were mapped to their respective KEGG and MetaCyc identifiers. This table is used in the Sco4Reconstruction script available from the Sco4 GitHub repository under ComplementaryScripts.<br> <b>S3 Table. Reactions selected from the <i>de novo </i>draft model for generating Sco4. </b>The merged <i>de novo </i>model (S1 Data) was manually evaluated to identified 398 reactions that constitute new pathways absent in iMK1208 and were subsequently added towards Sco4. This table is used in the Sco4Reconstruction script available from the Sco4 GitHub repository under ComplementaryScripts.<br> <b>S4 Table. Missing gene-associations in iMK1208 that were identified through <i>de novo</i> reconstruction. </b>Comparison of iMK1208 with the merged <i>de novo </i>reconstructions identified reactions that had no gene associated in iMK1208, while RAVEN 2.0 was able to identify the responsible genes. This table is used in the Sco4Reconstruction script available from the Sco4 GitHub repository under ComplementaryScripts.<br> <b>S5 Table. Spontaneous reactions included in Sco4, as identified from MetaCyc. </b>The <i>retrieveSpontaneous</i><i> </i>function was used to identify spontaneous reactions that connect to the Sco4<i> </i>model, as detailed in the Sco4Reconstruction script available from the Sco4 GitHub repository under ComplementaryScripts.<br> <b>S6 Table. Transport reactions added to Sco4 based on MetaCyc-based reconstruction and manual curation. </b>The merged <i>de novo </i>model (S1 Data) was manually evaluated to identified 10 transport reactions that were subsequently added to Sco4. This table is used in the Sco4Reconstruction script available from the Sco4 GitHub repository under ComplementaryScripts.<br> <b>S7 Table: Complete list of invalid KEGG identifiers and their replacements. </b>From iMK1208, 17 invalid KEGG metabolite identifiers were identified. Matching 13 of these metabolites to MetaCyc gave their valid KEGG identifier, leveraging on the annotation that is provided by MetaCyc. For the remaining 4 metabolites, it required matching their reaction between MetaCyc and KEGG to identify valid KEGG metabolite identifiers.<br> <b>S8 Table. Comparison of growth media validation results between iMK1208 and Sco4. </b>Culture and simulation conditions were as described previously (Kim <i>et al.</i>, 2014). Indicated are carbon and nitrogen sources that have been experimentally determined to be able to support growth of <i>S. coelicolor</i>, their exchange reaction, and whether growth is observed in Sco4 and iMK1208. Predictions on iMK1208 and Sco4 show identical results, indicating that the model expansion did not negatively affect model predictions.<br> <b>S9 Table: Identification of putative essential genes from mutagenesis study and predictions from the iMK1208 and Sco4 models. </b>Indicated is whether a gene has been identified as putative essential based on the Tn5 mutagenesis study (Xu <i>et al</i>., 2017) and simulation of gene essentiality in iMK1208 or Sco4.<br> <b>S10 Table. Combined FSEOF results on the secondary metabolites included in Sco4 model. </b>Target reactions identified from Sco4 for each of the indicated secondary metabolism product, as obtained from FSEOF analysis. Direction indicates whether a reaction is reversible and carries a flux during the FSEOF analysis in the forward direction (1), reverse direction (-1), or is irreversible (0).<br> <b>S11 Table. Definition of RAVEN model structure fields. </b>Where applicable, corresponding COBRA model structure fields are indicated.<br> <b>S12 Table. Matching of reaction identifiers between KEGG and MetaCyc. </b>Mapping is based on information provided by MetaCyc, while additional matches were identified by comparing reactants using the linkMetaCycKEGGRxns function.<b></b><br> <b>S13 Table. Matching of metabolite identifiers between KEGG and MetaCyc. </b>Mapping is based on information provided by MetaCyc.