Suppl Materials - UC San Diego Health Sciences
Contents
1. fraction of terms in each GO ontology that are covered by alignment to NeXO The overlap with GO in particular GO Cellular Components closely matches the results obtained with the original NeXO that included information from all four input networks an o T S Goo Additional GI oS x LL Additional GE Oo O X o O S U gc Z SS F z5 D 2 Q S c o PY E n O 5 O Z o S ue lt 0 10 20 30 40 50 100 500 2 0 10 20 30 40 50 100 500 Percent of interactions added Percent of interactions added C d E E Ea T Yo Yo wo i O O D N A N zo ZO 2 N g a A T o o 2o Gs ae To E gS ZS a oO Ww 0 10 20 30 40 50 100 500 s 0 10 20 30 40 50 100 500 U Y Percent of interactions added Percent of interactions added e f g h ToO TO T T x S on So O a 5 cite SO s Lo Len UN o 4 22 2 a 20 go T o ae bP OD S E zo z 3o DN 20 a Ee ES Eo Z BS g g 9 W a zZz oO O m o 13298 61627 Da 13298 61627 UO 13298 61627 UO 13298 61627 Number of PPIs Number of PPIs Number of PPIs Number of PPIs Supplementary Figure 2 Alignment results for NeXO using unfiltered networks The percent of NeXO terms that aligned to GO a and GO terms that aligned to NeXO b d when NeXO is built using more permissive interaction networks with 0 500 more genetic GI or co expression GE interactions The percent of NeXO terms that aligned to GO e and GO terms t2. provided separately NeXO USER MANUAL Navigation with Cytoscape The NeXO cys Supplementary File 1 is a Cytoscape session file that can be loaded directly into Cytoscape version 2 8 3 available from http www cytoscape org to allow for visualization and interactive analysis of the entire ontology The session provides three versions of NeXO the tree displaying the backbone of the ontology the DAG where supplementary term term relations are added and also the DAG with additional gene term assignments The following steps illustrate an exemplary analysis using NeXO cys and Cytoscape 1 Launch Cytoscape and use the File gt Open menu to locate and select the file NeXO cys 2 Select the NeXO Tree from the Network tab of the Control Panel at left below screenshot Cytoscape Desktop Session nexo_final_dag_final cys File Edit View Select Layout Plugins Help Tool eS Q Q q Q 1o o 3 Bes ta Search swi snf complex v lz lg Confifure Search Options for NeXO Tree Nexo Tree Network Ig vizMapper Eat ey wee r k Search at d 7 Ed gt N Network Nodes Edges cellular component CC Annotation root NexO DAG 9 Select Attribute Attribute Description cell E elie No description available part elles mi rotabule cytpskeleton Sample Attribute Values PSR1 YORO41 GAA GPl anchor transamidase ribonucleoprotein Je complex intracehwar prot aso me com
3. Supplementary Materials for A gene ontology inferred from molecular networks Janusz Dutkowski Michael Kramer Michal A Surma Rama Balakrishnan J Michael Cherry Nevan J Krogan amp Trey Ideker l 10 11 12 13 14 15 Supplementary Note 1 NeXO User Manual Navigation with Cytoscape and OBO Edit Supplementary Figure 1 Alignment results for NeXO without YeastNet and expression networks Supplementary Figure 2 Alignment results for NeXO using unfiltered networks Supplementary Figure 3 Alignment results using GO with excluded high throughput annotations Supplementary Figure 4 New NeXO term associated with the retromer complex Supplementary Figure 5 New term enriched for mutants sensitive to mercury chloride HgCl Supplementary Figure 6 Distinguishing is a and part of relations in NeXO Supplementary Figure 7 Representation of the proteasome with other clustering methods Supplementary Figure 8 Alignment results for NeXO using other hierarchical clustering methods Supplementary Figure 9 Alignment score threshold at 10 FDR Supplementary Figure 10 Distribution of robustness for aligned and unaligned terms Supplementary Table 1 Input networks investigated in this study Supplementary Tables 2 6 Excel spreadsheet files provided separately Supplementary File 1 NeXO in Cytoscape Format File NeXO cys provided separately Supplementary File 2 NeXO in Open Biomedical Ontology OBO Format File NeXO obo
4. es in the new term VPS8 VPS21 and the retromer subunits VPS17 and VPSS a b providing further support for the position of the new term next to the retromer complex in the NeXO ontology E mitochondial Mutant sensitivity art and to HgCl2 P value P g 10 e Normal Increased A Aa oxidored sctase activity YBR197C MTC2 MTC4 YPR153W antimonite trasport D MTC6 DLT1 Supplementary Figure 5 New term enriched for mutants sensitive to mercury chloride HgCl The term NeX0O 6375 is composed of six poorly uncharacterized ORFs MTC2 MTC4 MTC6 DLTI YBRI197C and YPRI53W NeXO places this terms under the mitochondrial component next to genes annotated with oxoreductase activity and antimonite transport Deletions of each of 6 genes in this term make yeast sensitive to mercury chloride HgCl P value 0 0018216 Mann Whitney U Test Descendant Connectivity Ratio is_arelation part_of relation Supplementary Figure 6 Distinguishing is a and part of relations in NeXO For each parent child relation in NeXO that aligned to a parent child relation in GO the parent was examined to compute its Child Connectivity Ratio CCR This was the ratio of the number of molecular interactions spanning different children of the parent to the total number of interactions falling within the parent CCR values for parent child relations in NeXO mapping to is a relations in GO are significantly lo
5. hat aligned to NeXO f h when NeXO is built using the original network and a more permissive interaction network which includes all 61627 physical protein protein interactions PPIs in BioGRID We observe that alignment results are robust across a wide range of more permissive input networks Note that in all cases the input networks contain PPI GI and GE data but exclude YeastNet to make sure that GO is not used for network training 0 35 0 30 0 25 0 20 0 15 0 10 0 05 0 00 S S A O g 0 6 0 5 0 4 0 3 0 2 0 1 0 0 No S S Kj E gS gS x NeXO terms aligned to GO Fraction GO CC terms aligned to NeXO Fraction gt we Se we oe G Ss lt S r K J es s s lt Ca MU g S LY lt S amp amp amp lt A x S IN ANEAN N we C d A T amp 0 25 S 0 25 U U X LL LL 0 20 0 20 gt x lt gt X lt 6 0 15 0 15 E v 5 0 10 5 0 10 a a 0 05 0 05 p oo 0 00 0 00 l x x x 2 R RA K we aY x Q amp Se x cS Xs ee KK KS X ee S K Y amp oS SO SX oe of K SN NI G Supplementary Figure 3 Alignment results using GO with excluded high throughput annotations The percent of NeXO terms that aligned to GO a and GO terms that aligned to NeXO b d when gene to term association based on high throughput interaction data HT and or those inferred from electronic annotations EA are removed from GO The specific evidence codes related to high throughpu
6. ing the backbone of the NeXO ontology X linked to Y indicates that X was linked to Y by an additional edge in the NeXO DAG File Edit Layout Editors Viewers Search Tools Metadata Reasoner Summary Config Help oue_0x if results all_text_fields contains proteasome 9 matches results all_text_fields contains swi snf 1 matches oF Ja results all_text_fields contains swi snf 1 matches ID Term Name Nex0 9301 ICC SWI SNF complex BP nucleosome mobilization MF positive regulation of transcription DNA dependent Ontology Tree Editor a child_of CC chromatin remodeling complex BP nucleosome organization MF A child_of CC BP chromatin remodeling MF a child_of Cc BP ATP dependent chromatin remodeling MF a child of Cc RSC complex BP nucleosome disassembly MF DNA dependent ATPase activity child of Cc BP chromatin remodeling at centromere MF child_of CC BP transcription elongation from RNA polymerase II promoter MF A child_of CC SWI SNF complex BP nucleosome mobilization MF positive regulation of transcription DNA dependent linked to CC BP chromatin remodeling at centromere MF linked to CC BP positive regulation of transcription involved in G1 S phase of mitotic cell cycle MF DNA binding esc child_of Cc BP termination of RNA polymerase transcription MF a Child of Cc BP histone modification MF histone H3 K4 meth
7. n complex assem child of CC replication fork BP DNA dependent DNA replication MF structure specific DNA binding shild of CC ribonucleoprotein complex BP ribonucleoprotein complex biogenesis MF structural molecule activity child of CC GO 0044391 BP ribosome biogenesis MF x lt j I il gt Definition Comment Cross Products Comment robustness 11 823186236341 density 1 bootstrap 0 82718052 cc_score 0 595589 bp_score 0 423599 mf _score 0 21875 Graph Editor ure_0x Set te CC methyl CC BP ATP dependent CC BP MFA histone complex chromatin nemodeling MF binding methylation hio Lore Xrefs Synonyms Subsets CC RSC complex BP SSS MINE TS BE BP regul nucleosome mobilization MF TA nucleosome disassembly a renia ion of recombine MF DNA dependent ATPase t p ipti aaa de NAD depenc activity nanscription ependen deacety o o child_of ce C BP positive negula J ranecrintian invoahsad AAG Commit Revert a NexO b GO Terms Aligned to NeXO Fraction GO CC GO BP GO MF Supplementary Figure 1 Alignment results for NeXO computed without the YeastNet and co expression networks a The Venn diagram shows the number of NeXO terms of size gt 4 that align to one or more GO ontologies colored circles or do not align white b The
8. on GO CC terms aligned to NeXO Fraction 0 00 0 00 C A 0 25 0 25 z gz 0 20 o 0 20 x lt x lt 2 2 o 0 15 o 0 15 ge ge Vv w 5 0 10 0 10 E 0 05 E 0 05 g g 3 Q O o V e 0 14 0 14 0 14 0 12 0 12 0 12 0 10 0 10 0 10 0 08 0 06 0 04 0 02 0 08 0 06 0 04 0 02 0 08 0 06 0 04 0 02 Jaccard score for NeXO and GO CC Jaccard score for NeXO and GO BP Jaccard score for NeXO and GO MF 0 00 0 00 0 00 gt xX R Ss A O O Ra g ge Supplementary Figure 8 Alignment results for NeXO using other hierarchical clustering methods Percent of NeXO terms that aligned significantly FDR 10 to any of the three GO ontologies a percent of terms in each of the GO ontologies that aligned to NeXO b d Jaccard index comparing the set of terms in NeXO to the set of GO Cellular Components e GO Biological Processes f and GO Molecular Functions g The original method achieves better agreement with GO than other standard methods ex Alignment against GO CC ex Alignment against GO BP ex Alignment against GO MF 2 3 4 5 8 gt 9 Term Size Number of Genes Alignment Score Threshold Corresponding to 10 FDR Supplementary Figure 9 Alignment score threshold at 10 FDR In the ontology alignment of NeXO to GO different alignment score thresholds were used for different NeXO term sizes in order to maintain a false discovery rate of 10 CC Cellular Component BP Biological Proce
9. plex site p of tS polarized growth me nuclear sae Ay Part chromatine 7 renfodeling v w complex ias spliteosomal complex Data Panel tG He Term or Gene Label Best Alignment Score cc Score Robustness SGD Gene Description 1 0 ALPHA 2 SUBUNIT OF THE 20S PROTEASOME 1 0 ALPHA 3 SUBUNIT OF THE 20S PROTEASOME THE ONLY NONESSEN 1 0 1 C ALPHA 1 SUBUNIT OF THE 20S PROTEASOME INVOLVED IN THE DEGR 0 570992 7 5 36285 None 1 0 1 0 ALPHA 6 SUBUNIT OF THE 20S PROTEASOME T gt l Node Attribute Browser Edge Attribute Browser Network Attribute Browser to Z00 Middle chck drag to PAN 3 Select the menu item View gt Show Graphics Details if the menu reads Hide Graphics Details this command is already active NeXO is visualized with high level terms labeled via alignment to the GO Cellular Component ontology Node size scales with the number of genes assigned at or below that term Node color intensity scales with the Best Alignment Score to GO See above screenshot and Fig 2 in the main text 4 Select parts of the tree using the mouse e g the proteasome subtree Information on selected nodes terms or genes are shown in the table in the Data Panel The main attributes columns are as follows Further details about the derivation of each attribute are provided in Supplementary Methods ID The NeXO identifier terms or SGD identifier genes of the node Term or Gene Label For terms name
10. s are transferred from GO alignment scores gt 0 1 CC Cellular Component BP Biological Process MF Molecular Function For genes common names are shown or if unavailable the ORF is shown CC Score The alignment score vs GO Cellular Component defaults to 1 0 for genes Best Alignment Score Maximum alignment score against the CC BP and MF ontologies of GO Robustness The term robustness score Supplementary Methods defaults to 1 0 for genes SGD Gene Description Provides the free form Gene Description field from SGD genes only Additional attributes such as term definitions CC BF MF term size number of assigned genes and interaction density can also be loaded using the Data Panel 5 Select the VizMapper tab of the Control Panel below screenshot and switch from the NeXO Overview to the NeXO CC Analysis visual style This display will color the nodes according to their alignment score against the GO Cellular Component ontology Green paths indicate consistent term term relations between the two ontologies 1 e ancestor descendant pairs that are aligned to GO terms that are also in an ancestor descendant relation In contrast a blue edge parent to child indicates a relation present in NeXO but absent from GO such that the GO terms aligned to the child and the nearest aligned NeXO ancestor are not in a descendant ancestor relation in GO Only NeXO terms with high alignment score CC score gt 0 2 are considered in thi
11. s relation analysis other terms are neutral for determining agreements and conflicts e g a green path may pass through them 6 Search NeXO for cellular components or genes of interest Use the mouse to push the button to the right of the Search field above screenshot to configure search options Select the CC Annotation attribute and push Apply Type swi snf into the search field and press Enter Gene names are also searchable using this attribute The SWI SNF complex will be shown 7 Select the NeXO DAG or NeXO DAG filtered from the Network tab of the Control Panel and repeat steps 3 6 In the full DAG version of the ontology edges are added which connect child terms to additional parents see below example in which a subunit of RSC is connected to the SWI SNF complex Similar analysis can also be performed using NeXO DAG with additional gene term assignments Cytoscape Desktop Session nexo_final_d j_final cys File Edit View Select Layout Plugins Help Tools BS QQ Q Q gi 3 2 E Search swi snf complex B E ontrol Panel i1 NeXO DAG Network Tg VizMapper Editor F ae gt Curren t Visual Style ame o 6 a one GINS oe comple T 2 chromatin amp zeta re A rasc complex a E e repair factor 1 complex o pn amp complex SWI SNF a s complex em nuclear chromatin ae repiisome accessibility ont complex de D pol
12. ss MF Molecular Function O m NeXO terms aligned to GO Cellular Components Oo o NeXO terms not aligned to GO Cellular Components gt oO AO Cc Y Q gt 2 O 6 a O AITITE i gt 0 5 2 5 4 5 6 5 8 5 10 12 14 16 18 20 22 24 26 Term Robustness Supplementary Figure 10 Distribution of robustness for aligned and unaligned terms The distributions are well separated for robustness gt 5 Supplementary Table 1 Input networks investigated in this study HQ used in all other analyses Network Type Used in Figs 1B C Number of Genes Number of Number of Number of Interactions Genes Interactions Physical Protein Protein 5 658 61 627 3 401 13 298 Co Expression 3 915 61 627 228 1 705 Genetic 4 396 61 627 3 090 11 240 YeastNet 5 268 61 627 3 351 10 573 Integrated High Quality HQ 5 051 29 789
13. t interaction data HT are IPI Inferred from Physical Interaction IEP Inferred from Expression Pattern and IGI Inferred from Genetic Interaction In the strict version we removed all associations of gene X to term Y if at least one of these associations met the condition for removal In the non strict version gene X could still be associated with term Y if it had at least one evidence code that did not meet the condition for removal NeXO alignment results appear robust to removal of HT and EA annotations from GO VPS21 VPS8 NeX0 8891 retromer VPS9 complex P E P 1 Genetic profile correlation 0 4 0 9 Oo a aa a a a a 25 LI lt Sa SA 6 4 t a O e Z O0 Te moO t n t NNANSTtOYt Ean N NDTORN o Co e FBS SY OF OU SAYS TOO LOPS SEY USER SOS SV SSE ZZOU YE ELSYENS OS SY VeENyeCuguNSxXy AEE lt xrs gt tazZzaa Zaz n EzI co iaTaanvusvraertan axeuwmasziic eee erALHOLTUt sa te OUOU RUS eS SOS Ze OU SSG OUSSUEESSSU SAwWAS gt oO lt uw VPS21 VPS8 VPS17 VPS5 SYN8 VPS8 3 2 1 O 1 2 3 negative interactions positive interactions synthetically sick lethal suppressive epistatic Supplementary Figure 4 New NeXO term associated with the retromer complex New term NeX0O 8891 composed of VPS8 VPS21 and VPS9 and placed directly next to the retromer complex in NeXO a The new genetic interaction screen identified very high profile correlation between gen
14. wer than for NeXO relations mapping to part_of relations e 5 e e particle regulatory ka CORE COMPLEX AT eP picie bg ALPHA SUBUNIT su complex o gubcomplex ol a regulatory COM Pies partitle beta subunit core complex alpha subunit REGULATORY PARTICLE LID COMPLEX core somelex om cunt complex regulator 5 cout y regilatorye s stora 2 ba a p rticle subcompiex e core regulatory Mi R z o 7 beta subunit suusbmplex core complex e proteasome plexo proteasome complex ore re la ory e e complex parti AT regula ery TESE articl o p e p the K 4 P astic ran a enue suei Subgpmp ex core L tomplex beta subunit core complex bd alpha eubunite ee o Supplementary Figure 7 Representation of the proteasome when using other clustering methods in the NeXO approach original a edge betweenness b UPGMA based on topological overlap distance c and UPGMA based on commute distance d NeXO alignment is performed to map the resulting trees to the GO Cellular Component ontology The alternative clustering methods b d do not recover the hierarchy to the same extent as the original method used by NeXO a While core and regulatory particles are correctly identified organization of more specific components is incorrect b d 0 35 0 6 0 30 0 25 ee 0 4 0 3 0 2 0 20 0 15 0 10 0 05 0 00 0 0 NeXO terms aligned to GO Fracti
15. y com Edge Tooltip mn keg Lj av B 3 EB fe amp bs 9 ID Colo Node Border Of Term or Gene Label Best Alignment Score cc Score Robustness SGD Gene Description Node Attribute Browser Edge Attribute Browser Welcome to Cytoscape 2 8 3 Right click drag to ZOOM Middie click drag to PAN Network Attribute Browser Navigation with OBO Edit The NeXO ontology can also be explored using designated ontology editors and browsers such as OBO Edit available from http www oboedit org as shown in the below screenshot For this purpose we have provided the file NeXO obo Supplementary File 2 which conforms to the Open Biological and Biomedical Ontology OBO format Using OBO Edit the NeXO ontology can be explored in three basic ways e Using the Ontology Tree Editor e Using the Graph Editor e By searching for keywords appearing in NeXO term annotations Terms in the NeXO obo file are annotated with text labels transferred through alignment to GO The label indicates matchings with score gt 0 1 to any one of the three GO ontologies Cellular Component CC Biological Process BP and Molecular Function MF Additional information about the term such as its robustness score interaction density and alignment scores are provided in the Comment field Each term relationship is annotated with one of two types child of or linked to The X child of Y relationship indicates that X is a child of Y in the tree form
16. ylation a child of CC BP MF histone binding a child_of CC methyltransferase complex BP histone methylation MF histone methylation A hild of Cc nuclear chromatin BP MF Sl child of cc BP MF a child_of Cc BP MF 2 child of CC BP MF child_of CC BP cytoplasmic sequestering of protein MF child of Cc BP protein sumoylation MF linked to CC DNA helicase complex BP vesicle transport along actin filament MF 3 5 DNA helicase activity child of CC Ino80 complex BP MF DNA end binding a4 child sof CC BP MF child_of CC BP MF child_of CC BP MF OHA Edit liked to CC DNA helicase complex BP vesicle transport along actin filament MF 3 5 DNA helicase activity Text Editor lifked to CC DNA helicase complex BP vesicle transport along actin filament MF 3 5 DNA helicase activity ID Nex0 9301 Child_of CC H4 H2A histone acetyltransferase complex BP MF Namespace NexO child of CC DNA helicase complex BP vesicle transport along actin filament MF 3 5 DNA helicase activity P A Child of CC Rpd3L Expanded complex BP negative regulation of chromatin silencing at telomere MF histone deacetyla Name CC SWI SNF complex BP nucleosome mobilization MF positive regulation of transcription DNA dependent child_of CC DNA directed RNA polymerase II holoenzyme BP RNA polymerase II transcriptional preinitiatio
Download Pdf Manuals
Related Search