Computer Laboratory University of Cam bridge CytoCom: a
Contents
1. www cl cam ac uk mam211 In Cytoscape 3 1 0 or later version the App Manager allows users to quickly install and uninstall the apps After downloading CytoCom please go to Apps app Manager and then click the Install from File button at the bottom of the Install Apps tab Select the downloaded CytoCom jar to install After installing CytoCom it could be found under the Control Panel Documentation Understanding the comorbidity of human diseases is one of the most chal lenging issues in bioinformatics today CytoCom is a plug in for Cytoscape to visualise query and analyse disease disease networks Data set We collected statistically significant pairwise comorbidity associations re constructed from over 32 million medical records in the US Medicare claims database recorded in the ICD 9 format http www icd9data com which are frequently used for epidemiological and demographic studies and collected from Hidalgo et al 2009 We used MedPAR records from 1990 to 1993 where the dates and reasons for all hospitalisations were reported in ICD 9 CM format and it contains the diagnoses of 13 039 018 elderly pa tients In total the ICD 9 classification consists of 657 different categories at the 3 digit level Hidalgo et al 2009 Yet the data set is large enough to predicate race and gender specific comorbidity patterns Control Panel After launching Cytoscape CytoCom will appear in the control panel an2. Computer Laboratory a O0 TD a nsi CO G gt gt A D Q Es Computer Laboratory University of Cambridge Cambridge UK CB3 0FD User Manual CytoCom a Cytoscape plugin to visualise query and analyse disease disease dynamic networks CytoCom Version 1 0 0 Haoming Xu Mohammad Ali Moni and Pietro Li October 2014 User Manual CytoCom a Cytoscape plugin to visualise query and analyse disease disease dynamic networks CytoCom Version 1 0 0 Haoming Xu Mohammad Ali Moni and Pietro Lid Computer Laboratory University of Cambridge Cambridge UK CB3 0FD October 2014 Table of Contents System Requirements sise iii TETTE IMEEM NEIN iv Eu iv Control Panel 2 ble lE Rie a cibi ko den Sud aa a b anan a a a gana iv Parameter settingl se nnns hr ehh sessi essa vii Network construction ccc cece ccc c cece n RR eem meme ease e esta e esee vii Network Visualisation nn I hme e eme emen a gg a ease ena viii Customise the network aaa eaaa cece cence eee Ree he ese ee semen astra viii References iii System Requirements In order to use CytoCom your computer must be equipped with the follow ing software package e Java 1 6 is needed Java 1 7 is recommended e Cytoscape 3 1 0 installed To download CytoCom please visit http
3. ES AND IMMUNITY DISORDERS MENTAL DISORDERS DISEASES OF THE NERVOUS SYSTEM AND SENSE ORGANS SYMPTOMS SIGNS AND ILL DEFINED CONDITIONS NEOPLASMS CERTAIN CONDITIONS ORIGINATING IN THE PERINATAL PERIOD DISEASES OF THE RESPIRATORY SYSTEM DISEASES OF THE SKIN AND SUBCUTANEOUS TISSUE INFECTIOUS AND PARASITIC DISEASES DISEASES OF THE CIRCULATORY SYSTEM DISEASES OF THE DIGESTIVE SYSTEM INJURY AND POISONING CONGENITAL ANOMALIES DISEASES OF THE MUSCULOSKELETAL SYSTEM AND CONNECTIVE TISSUE DISEASES OF THE GENITOURINARY SYSTEM DISEASES OF THE BLOOD AND BLOOD FORMING ORGANS COMPLICATIONS OF PREGNANCY CHILDBIRTH AND THE PUERPERIUM Figure 5 Legend window menu it is possible to apply a specific layout algorithm to generate a partic ular view according to the user s preferred choice In addition the user can change the size of the existing network nodes and the width of the existing network edges by using the Style properties of the Cytoscape xi References Hidalgo C A Blumm N Barab si A L and Christakis N A 2009 A dynamic net work approach for the study of human phenotypes PLoS computational biology 5 4 e1000353
4. cted patients from the specific pop ulation Columns rr and phi are two statistical measures CytoCom built disease comorbidity network based on these information File Edit View Select Layout Apps Tools Help m mie sd QQ QQ s Control Panel Tg Network style select E Cytocom Dataset Gender Female Both Race Inputs im Console INFO Loading ICD 9 3 digit datasets INFO Loaded INFO Explore the network with 350 INFO Explore the network with 966 INFO Explore the network with 344 INFO Explore the network with 741 WARNING 341 can t explore the network Table Panel INFO Explore the network with 507 WARNING 507 has already been explored x DD ge oo B Tuy GO f x Comorbidity Network 340 0 01 3 disease 1 name 1 prevalence 1 disease 2 name2 phi 344 Other paralytic syndromes 87151 Multiple sclerosis 15936 0 070353 350 Trigeminal nerve disorders 18537 Multiple sderosis 15936 0 026877 596 Other disorders of bladder 359673 Multiple sderosis_ 15936 0 016410 599 Other disorders of urethra and urinary tract 2404493 Multiple sderosis_ 15936 0 024742 707 Chronic ulcer of skin 442019 Multiple sderosis_ 15936 0 021891 821 Fracture of other and unspecified parts of 80104 Multiple sderosis 15936 0 010228 351 Facial nerve disorders 24365 Trigeminal ner 18537 0 017832 352 Disorders of oth
5. d will present itself as shown in Figure 2 In order to build a disease disease association network we need to load the data according to the following steps 1 Gender selection Select the Male Female or Both options from the database 2 Race selection Select the White Black or Both options from the database 3 Click the Load button the CytoCom will be connected to the inner database from which the data will be loaded After loading the data we can explore disease disease association comorbid ity network by means of the following steps Male Female Both White Black Both INFO Loading ICD 9 3 digit datasets INFO Loaded INFO Explore the network with 707 Figure 1 CytoCom control panel vi Enter the phi values in the text fields It could be either integer or dec imal number Hence values greater than input value will be selected to explore the network 2 Enter the relative risk values in the text fields It could be either integer or decimal number Hence values greater than input value will be se lected to explore the network 3 Click the Apply button The CytoCom will then construct the disease disease network diagram The name of the network starts with Comor bidity network ICD codes Phi RR See Figure 2 sa Comorbidity Network 340_0 01_3 8 i Comorbidity Network 340 0 01 3 8 60 0 3140 Figur
6. e 2 The name of the network is Comorbidity network 340 0 01 0 6 vii In order to explore more diseases that are associated with the disease nodes as seed input in an existing network users need to search the network by double clicking the selected nodes within it A search is then performed based on these nodes using the pre specified Phi and RR values and the additional comorbid diseases are then added to the existing network This function works with one or more selected nodes and supports the creation of dynamic network diagrams interactively with CytoCom Figure 4a shows the network generated with CytoCom using disease 340 and Figure 4b shows the extended disease disease association network starting from the input disease code 350 Parameter setting For a pair of diseases and 7 we used two statistical measures to quantify the relationship between two diseases Relative Risk RR and correlation ij which are calculated based on the Hidalgo et al 2009 The corre lation of RRi 1 implies no comorbidity RR gt 1 implies positive co morbidity and 0 lt RR lt 1 implies negative comorbidity Similarly Qij 0 implies no co morbidity 0 lt lt 1 implies positive comorbid ity and 1 lt lt 0 implies negative comorbidity The two comorbidity measures are not completely independent of each other and both measures have their intrinsic biases 2009 They increase with the num ber of patients a
7. er cranial nerves 3187 Trigeminal ner 18537 0 014123 Node Table Edge Table Network Table Network Visualisation Each node in the network represents a unique disease Two diseases are con nected if there is an association between them We colour each node accord ing to the category of diseases based on the first 3 digits of the given ICD9 codes The node size increases with the increasing the disease prevalence More common diseases are represented as larger nodes as shown in Figure In addition the user can retrieve a display of the legend by clicking the Show legend button It opens the Legend window which provides colour information for each node as shown in Figure 5 All nodes can be labelled with either the ICD 9 CM code or the disease name and the Show names or Show codes tabs can be clicked alternately to switch between labelling the nodes by code or by name Customise the network Cytoscape provides several layout algorithms for organising its network vi sualisation system For example by selecting Layouts in the Cytoscape a b Figure 4 Panels a and b show the disease disease network Panel a shows the network generated with CytoCom using disease 340 and Panel b shows the diseases associated with 344 in the existing network that is expended 0080008000800 06 00 i ENDOCRINE NUTRITIONAL AND METABOLIC DISEAS
8. ffected by both diseases For example RR overestimates relationships involving among rare diseases and underestimates the comor bidity between highly prevalent illnesses whereas o accurately discriminates comorbidities between pairs of diseases of similar prevalence but underes timates the comorbidity between rare and common diseases Hidalgo et al suggested that two diseases are strongly associated if R gt 20 and Qij gt 0 06 Therefore user may consider these values for relative risk and to estimate the comorbity among diseases However user may observer the disease associations by putting their desired parameters values Network construction The comorbidity disease network is constructed using the two comorbidity measures relative risk and phi correlation and using the patient medical records All informations that are provided as a data source file is used for the identification and calculation of the comorbidity association between dis eases according to the parameter setting by the users The table panel in Figure 3 provides the representation of the comorbidity output network in formation Table contains 8 columns disease 1 and disease 2 headed columns indicate ICD 9 code of disease 1 and disease 2 respectively Column name 1 viii and name 2 represent names of disease 1 and disease 2 Column prevalence 1 and prevalence 2 mean the prevalence of diseases Both of the prevalence values are the absolute number of affe
Download Pdf Manuals
Related Search