LanDis: The Disease Landscape Explorer

LanDis: The Disease Landscape Explorer | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Brief Communication LanDis: The Disease Landscape Explorer Alberto Paccanaro, Horacio Caniza, Juan Cáceres, Mateo Torres This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3168447/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 10 Jan, 2024 Read the published version in European Journal of Human Genetics → Version 1 posted 10 You are reading this latest preprint version Abstract From a network medicine perspective, a disease is the consequence of perturbations on the interactome. These perturbations tend to appear in a specific neighbourhood on the interactome, the disease module, and modules related to phenotypically similar diseases tend to be located in close-by regions. We present LanDis, a freely available web-based interactive tool (paccanarolab.org/landis) that allows domain experts, medical doctors and the larger scientific community to graphically navigate the interactome distances between the modules of over 44 million pairs of heritable diseases. The map-like interface provides detailed comparisons between pairs of diseases together with supporting evidence. Every disease in LanDis is linked to relevant entries in OMIM and UniProt, providing a starting point for in-depth analysis and an opportunity for novel insight into the aetiology of diseases as well as differential diagnosis. Biological sciences/Computational biology and bioinformatics/Genome informatics Health sciences/Medical research/Genetics research Biological sciences/Computational biology and bioinformatics/Data processing Disease similarity genetic diseases systems medicine network medicine differential diagnosis Figures Figure 1 Figure 2 INTRODUCTION In recent decades, our understanding of diseases and their causes has shifted from simple relationships between genes and diseases to more comprehensive models, which take into account the interplay of disease genes through their multiple molecular interactions. Studying diseases in the context of the human interactome has revealed that a disease’s causal genes tend to cluster in close-by regions – the disease module – and that diseases that share causal genes tend to exhibit phenotypical similarity (1). The idea that closeness on the interactome relates to phenotypical similarity has applications in disease gene prediction and differential diagnosis (1, 6-8). For instance, recent methods have successfully exploited these concepts to prioritise candidate disease genes according to their level of connectivity to known disease genes (2-7). Moreover, the comprehensive study of the phenotypical similarities of diseases can help in understanding their aetiology and reveal commonalities in their pathophysiology. A few measures have been developed to systematically quantify the similarity between pairs of diseases (See Supplementary Note 1). LanDis relies on the Caniza measure, which summarises the information about diseases that is scattered across the biomedical literature (8). The method is based on the idea that a disease can be described accurately by the set of MeSH terms used to annotate the publications relevant for that disease. Pairwise similarities between diseases are then calculated by exploiting the structure of the MeSH ontology. A comparison of the different similarity measures using sets of diseases with known disease genes, showed that the Caniza similarity outperforms all other measures in terms of accuracy at predicting closeness of disease modules on the interactome (8). This is probably due to the large volume of information, i.e. the thousands of disease related publications, which contribute to the measure. While the importance of disease similarity measures for medical research is clearly understood, until now their use in practice has been limited. An important reason is that disease similarities are mainly available only as matrices containing millions of numerical values, one for each disease pairs, and this limits the scientists’ ability to use this information for reasoning and making inferences. In this paper we present LanDis, a freely available web server that provides an intuitive interface to analyse millions of similarity relationships between heritable diseases, together with the evidence supporting such relationships. RESULTS In LanDis, the similarity landscape is represented as a graph in which nodes are diseases and links are labelled with the Caniza similarity score between the diseases they connect. Figure 1 shows the landscape of the OMIM disease Tetralogy of Fallot , TOF (MIM: 187500), represented by the central node in the figure. TOF is a congenital heart defect characterised by a ventricular septal defect, pulmonary valve stenosis, thickened right ventricle and overriding aorta (9). Patients with TOF develop cyanosis in proportion to the pulmonary valve stenosis, rapid breathing to compensate low oxygen levels and a heart murmur. Let us analyse each disease that we find connected to TOF in our similarity landscape. The Conotruncal Heart Malformations CHTM (MIM: 217095) disorder includes the TOF malformations and is known to be causally related to gene NKX2-5, a gene also known to be causally related to TOF. Both Alagille Syndrome 1 ALGS1 (MIM:118450) and Right Atrial Isomerism RAI (MIM:208530) not only share phenotypic similarities with TOF such as pulmonary stenosis (ALGS1) and complete atrioventricular septal defects (RAI), but also have disease genes in common with TOF, namely JAG1 and GDF1 (10). Congenital heart defects, Multiple Types CHTD6 (MIM: 613854) (formerly Transposition of the great arteries DTGA3) often have ventricular septal defects and associations between CHTD6 and the TOF-associated gene GDF1 have been reported in the literature (11). Aortic Arch Interruption, Facial Palsy, Retinal Coloboma (MIM: 107550) exhibits symptomatic similarities with TOF, such as fatigue, rapid breathing, fast heart rate, low oxygen levels among others (12). As is the case with TOF, the aortic arch interruption is characterised (among other features) by a ventricular septal defect. Finally, Takayasu Arteritis (MIM: 207600) is an inflammatory disease of the arteries, with predilection for the aorta and its branches. The disease is characterised by lesions that can, among others, have stenotic qualities (13). Interestingly, the diseases in the graph without a direct connection to TOF reflect not only their associations with their immediate neighbours but also, to some extent, with TOF. For example, DiGeorge syndrome DGS (MIM: 188400) not only shares a gene with TOF (TBX1), but also the outflow tract defects present in DGS are associated with a higher incidence of conotruncal abnormalities (14). LanDis is a web application in which the user can interact with all the elements in the graph and the diseases can be repositioned either by dragging them or through several pre-defined layouts (circular, concentric, grid, breadth-first and force directed). Seamless exploration of the diseases similarity landscapes can be performed through the selection of any disease in the landscape. Every disease similarity landscape can be downloaded in publication-quality, high-resolution PNG images for offline analysis. Users can also select a disease and obtain a catalogue of those diseases most similar to it in a tabular format, as well as a detailed comparison between pairs of diseases – Figure 2 shows the Compare page for TOF and ALGS1. For users who wish to use the Caniza similarity data as part of a larger pipeline, a CSV plain-text file is available from the download section of the website. To ease further exploration, LanDis links every MeSH term, disease and disease gene to its corresponding entry in the OMIM, UniProt and National Library of Medicine websites respectively. DISCUSSION LanDis offers a new perspective to explore disease similarity relationships. It is a simple and powerful tool which can be used for differential diagnosis as diseases that present similar molecular features will be assigned high similarity. Importantly, LanDis provides the user with a rationale for the results by making available the set of MeSH terms, corresponding to disease phenotypes, used to calculate the disease similarity. In this way, scientists can focus on the clinical features deemed more critical while concentrating on a selected list of highly similar diseases. Notably, LanDis is able to find similarities at molecular level between diseases even in the absence of any molecular information – this is because it only needs a list of publications associated with each disease. Supplementary Figure 1 shows the number of publications, MeSH terms and genes associated to the diseases in LanDis. As is expected, a disease with many referenced publications tends to be annotated by many MeSH terms, but a high number of publications does not necessarily correspond to a high number of known genes – for example, Huntington’s disease, that has more than 450 references and close to a 1000 MeSH terms, is associated to a single gene. However, since LanDis relies exclusively on publications and their corresponding MeSH terms, the sparseness of molecular information does not prevent the similarity scores from being calculated. In fact, LanDis attempts to encapsulate all available information about diseases – for example, the references of type 2 Diabetes (NIDDM) include information about several clinical trials and multi-year studies on the effects of glucose on insulin levels. LanDis aims at becoming a support tool for bioinformaticians as well as medical practitioners. It is freely available through its website, no registration or installation is needed and our servers store no information about the users. ONLINE METHODS Disease similarities and datasets LanDis mines OMIM to extract 139,549 PubMed references. For each publication, LanDis queries the Medline API obtaining a total of 17,110 MeSH terms. A few disease entries in OMIM with no references or MeSH annotations are excluded from LanDis, for a working total of 9,735 diseases. This amounts to over 44.7 million similarities, one per disease pair. To produce the pairwise similarities, LanDis relies on the structure of the MeSH ontologies. The similarity between a pair of diseases is given by the Resnik similarity of the sets of MeSH terms annotating the diseases (16). The Resnik similarity score of two sets of MeSH terms is given by the information content of their lowest common ancestor, which is defined as the negative logarithm of the probability of finding it among the annotations of the OMIM diseases (16-18). MeSH terms are organised into 16 ontologies and a given disease can be annotated with terms from more than one ontology. This means that for every disease up to 16 similarities can be calculated. Following Caniza et al. (8), LanDis exploits the fact that these ontologies are interconnected to combine them and produce a single score. Implementation details LanDis is implemented using Python and the Django framework, following a strict Model-View-Controller architecture. The data persistence is provided by a single-file SQLite database, which holds the similarity data and all additional information required to provide LanDis functionalities. Indices where defined to improve access time to the SQL database. The user interface was designed using HTML 5 and the JQuery JavaScript library. Additionally, two well-known JavaScript libraries, D3.js and Cytoscape.js, are included. D3.js provides the tools for dynamic visualisations of the similarity data and Cytoscape.js provides the engine for LanDis disease landscape explorer. This allows for a flexible interface that fits most resolutions for desktops, laptops and most mobile devices. There are no special requirements for a user's computer, since all user-side JavaScript code was carefully developed to reduce its footprint. Warnings are displayed for larger more resource-consuming plots, allowing the user to choose whether to continue with the operation. The source code is freely available from GitHub at https://github.com/paccanarolab/landis and is released under the GPLv3 license. We have tested LanDis on all major browsers and operating systems (mobile and desktop), and it performs best on Google Chrome. DECLARATIONS Dr. Caniza has nothing to disclose. Dr. Cáceres has nothing to disclose. Dr. Torres has nothing to disclose. Dr. Paccanaro has nothing to disclose. FUNDING: A.P. was supported by Biotechnology and Biological Sciences Research Council (https://bbsrc.ukri.org/) grant numbers BB/K004131/1, BB/F00964X/1,and BB/M025047/1; Medical Research Council (https://mrc.ukri.org) grant number MR/T001070/1; Consejo Nacional de Ciencia y Tecnología Paraguay (https://www.conacyt.gov.py/) grants numbers 14-INV-088 (to AP, JC, MT and HC), PINV15–315, and PINV20-337; National Science Foundation Advances in Bio Informatics (https://www.nsf.gov/) grant number 1660648; Fundacão de Amparo a Pesquisa do Estado do Rio de Janeiro (https://www.faperj.br) grant number E-26/201.079/2021 (260380); Conselho Nacional de Desenvolvimento Científico e Tecnológico (https://www.cnpq.br) grant number 311181/2022-8; and Fundacão Getulio Vargas. Authors' contributions: HC and AP developed the model. HC conceptualised the software. HC, JC and MT developed LanDis. AP tested and provided feedback on the features of LanDis. HC and AP wrote the manuscript. All authors read and approved the final manuscript. Acknowledgements: We thank Diego Galeano for useful discussions on the user interface. References Barabási A-L, Gulbahce N, Loscalzo J. Network medicine: a network-based approach to human disease. Nature Reviews Genetics. 2011;12(1):56-68. Xiujuan Wang NGaHY. Network-based methods for human disease gene prediction Briefings in Functional Genomics. 2011:280–93. Zou Q, Li J, Wang C, Zeng X. Approaches for recognizing disease genes based on network. BioMed research international. 2014;2014. Zou Q, Li J, Song L, Zeng X, Wang G. Similarity computation strategies in the microRNA-disease network: a survey. Briefings in functional genomics. 2015;15(1):55-64. Zou Q, Li J, Hong Q, Lin Z, Wu Y, Shi H, et al. Prediction of microRNA-disease associations based on social network analysis methods. BioMed research international. 2015;2015. Gliozzo, J., Perlasca, P., Mesiti, M., Casiraghi, E., Vallacchi, V., Vergani, E., ... & Valentini, G. (2020). Network modeling of patients' biomolecular profiles for clinical phenotype/outcome prediction. Scientific Reports, 10(1), 3612. Cáceres, J. J., & Paccanaro, A. (2019). Disease gene prediction for molecularly uncharacterized diseases. PLoS computational biology, 15(7), e1007078. Caniza H, Romero AE, Paccanaro A. A network medicine approach to quantify distance between hereditary disease modules on the interactome. Scientific reports. 2015;5. OMIM. OMIM Entry 187500. Online Mendelian Inheritance in Man2017. Gruber PJ, Epstein JA. Development Gone Awry. Circulation Research. 2004;94:273-83. Karkera JD, Lee JS, Roessler E, Banerjee-Basu S, Ouspenskaia MV, Mez J, et al. Loss-of-Function Mutations in Growth Differentiation Factor-1 (GDF1) Are Associated with Congenital Heart Defects in Humans. American journal of human genetics. 2007:81 (5): 987-94. Collins-Nakai RL, Dick, M., Parisi-Buckley, L., Fyler, D. C., & Castaneda, A. R. Interrupted aortic arch in infancy. The Journal of pediatrics. 1976:88(6), 959-62. Saruhan-Direskeneli G, Hughes, T., Aksu, K., Keser, G., Coit, P., Aydin, S. Z., ... & Hoffman, G. S. Identification of multiple genetic susceptibility loci in Takayasu arteritis. . The American Journal of Human Genetics,. 2013: 93(2), 298-305. Bruneau BG. The developmental genetics of congenital heart disease. Nature. 2008:451(7181), 943. McCright B, Lozier, J., & Gridley, T. A mouse model of Alagille syndrome: Notch2 as a genetic modifier of Jag1 haploinsufficiency. Development. 2002:129(4), 1075-82. Resnik, P. Semantic Similarity in a Taxonomy: An Information-Based Measure and its Applications to Problems of Ambiguity in Natural Language. Journal of Artificial Intelligence Research 11 (1999) 95-130. Yang H., Nepusz T., Paccanaro A. Improving GO semantic similarity measures by exploring the ontology beneath the terms and modelling uncertainty, Bioinformatics, Volume 28, Issue 10, May 2012, Pages 1383–1389. 18. Caniza, H., Romero, A. E., Heron, S., Yang, H., Devoto, A., Frasca, M., ... & Paccanaro, A. (2014). GOssTo: a stand-alone application and a web tool for calculating semantic similarities on the Gene Ontology. Bioinformatics, 30(15), 2235-2236. Additional Declarations There is no duality of interest Supplementary Files SupplementaryInformation.docx Supplementary Material Cite Share Download PDF Status: Published Journal Publication published 10 Jan, 2024 Read the published version in European Journal of Human Genetics → Version 1 posted Editorial decision: revise 03 Oct, 2023 Review # 2 received at journal 01 Sep, 2023 Reviewer # 2 agreed at journal 01 Sep, 2023 Review # 1 received at journal 23 Aug, 2023 Reviewer # 1 agreed at journal 23 Aug, 2023 Reviewers invited by journal 22 Aug, 2023 Submission checks completed at journal 24 Jul, 2023 First submitted to journal 20 Jul, 2023 Unknown event 17 Jul, 2023 Editor assigned by journal 13 Jul, 2023 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3168447","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Brief Communication","associatedPublications":[],"authors":[{"id":228322890,"identity":"86bf5c75-7193-442b-8ae9-81f49fcfa072","order_by":0,"name":"Alberto Paccanaro","email":"data:image/png;base64,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","orcid":"","institution":"Royal Holloway, University of London","correspondingAuthor":true,"prefix":"","firstName":"Alberto","middleName":"","lastName":"Paccanaro","suffix":""},{"id":228322891,"identity":"3fabbc8a-ee0b-4442-a805-d26ad083c8a2","order_by":1,"name":"Horacio Caniza","email":"","orcid":"","institution":"Universidad Paraguayo Alemana","correspondingAuthor":false,"prefix":"","firstName":"Horacio","middleName":"","lastName":"Caniza","suffix":""},{"id":228322892,"identity":"785d46c5-9c1a-4dad-8266-c9c8380c5546","order_by":2,"name":"Juan Cáceres","email":"","orcid":"","institution":"Royal Holloway, University of London","correspondingAuthor":false,"prefix":"","firstName":"Juan","middleName":"","lastName":"Cáceres","suffix":""},{"id":228322893,"identity":"aa3d7822-84a0-48ab-88fc-11c5867faeb2","order_by":3,"name":"Mateo Torres","email":"","orcid":"","institution":"Fundação Getúlio Vargas","correspondingAuthor":false,"prefix":"","firstName":"Mateo","middleName":"","lastName":"Torres","suffix":""}],"badges":[],"createdAt":"2023-07-13 21:11:44","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3168447/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3168447/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41431-023-01511-9","type":"published","date":"2024-01-10T05:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":42245886,"identity":"0181aec5-9076-44c8-8c96-e884140397e6","added_by":"auto","created_at":"2023-08-28 14:19:27","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":100055,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eThe disease similarity landscape of the congenital heart defect Tetralogy of Fallot, TOF.\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e Each node represents a disease and links are coloured based on the Caniza similarity between the linked diseases. The diseases in the TOF landscape are either directly associated to heart conditions, such as the case of ALGS1 (15), CTHD6 (11) and RAI or indirectly through some common phenotypic features such as DGS\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3168447/v1/d261062798d1b8ccc28585a1.jpg"},{"id":42244327,"identity":"7cc18b3f-333a-4dc7-a30b-fe9f4049ea03","added_by":"auto","created_at":"2023-08-28 14:11:27","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":85790,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eDetailed comparison of TOF and ALGS1\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e. In the histogram, each bar represents the percentage of disease pairs with similarity score in the corresponding range. The red circle indicates the range of the similarity of the TOF – AGS1 pair. The bottom of the figure shows the MeSH annotations grouped by ontology – these are accessible by expanding the collapsed sections. Links to the relevant MeSH term pages and OMIM pages are also available. The known disease genes are linked to their UniProt page.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3168447/v1/eacc09d9faba7a54b14d7a35.jpg"},{"id":49992751,"identity":"129dcf74-30e5-4403-b5ef-c650babd99b6","added_by":"auto","created_at":"2024-01-22 18:58:49","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":356549,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3168447/v1/8e6be1cf-265c-4e62-aed7-b41a9d2696af.pdf"},{"id":42244325,"identity":"9d49358a-03b4-4555-ab6e-8ebcfef0d21d","added_by":"auto","created_at":"2023-08-28 14:11:27","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":72989,"visible":true,"origin":"","legend":"Supplementary Material","description":"","filename":"SupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-3168447/v1/4973ba1388e7c499f5bcc36e.docx"}],"financialInterests":"There is no duality of interest","formattedTitle":"LanDis: The Disease Landscape Explorer","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eIn recent decades, our understanding of diseases and their causes has shifted from simple relationships between genes and diseases to more comprehensive models, which take into account the interplay of disease genes through their multiple molecular interactions. Studying diseases in the context of the human interactome has revealed that a disease\u0026rsquo;s causal genes tend to cluster in close-by regions \u0026ndash; the disease module \u0026ndash; and that diseases that share causal genes tend to exhibit phenotypical similarity\u0026nbsp;(1). \u0026nbsp;The idea that closeness on the interactome relates to phenotypical similarity has applications in disease gene prediction and differential diagnosis (1, 6-8). For instance, recent methods have successfully exploited these concepts to prioritise candidate disease genes according to their level of connectivity to known disease genes\u0026nbsp;(2-7). \u0026nbsp;Moreover, the comprehensive study of the phenotypical similarities of diseases can help in understanding their aetiology and reveal commonalities in their pathophysiology.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA few measures have been developed to systematically quantify the similarity between pairs of diseases (See Supplementary Note 1). LanDis relies on the Caniza measure, which summarises the information about diseases that is scattered across the biomedical literature (8). The method is based on the idea that a disease can be described accurately by the set of MeSH terms used to annotate the publications relevant for that disease. Pairwise similarities between diseases are then calculated by exploiting the structure of the MeSH ontology. A comparison of the different similarity measures using sets of diseases with known disease genes, showed that the Caniza similarity outperforms all other measures in terms of accuracy at predicting closeness of disease modules on the interactome (8). This is probably due to the large volume of information, i.e. the thousands of disease related publications, which contribute to the measure.\u003c/p\u003e\n\u003cp\u003eWhile the importance of disease similarity measures for medical research is clearly understood, until now their use in practice has been limited. An important reason is that disease similarities are mainly available only as matrices containing millions of numerical values, one for each disease pairs, and this limits the scientists\u0026rsquo; ability to use this information for reasoning and making inferences.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn this paper we present LanDis, a freely available web server that provides an intuitive interface to analyse millions of similarity relationships between heritable diseases, together with the evidence supporting such relationships.\u0026nbsp;\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cp\u003eIn LanDis, the similarity landscape is represented as a graph in which nodes are diseases and links are labelled with the Caniza similarity score between the diseases they connect. \u003cstrong\u003eFigure 1\u0026nbsp;\u003c/strong\u003eshows the landscape of the OMIM disease \u003cem\u003eTetralogy of Fallot\u003c/em\u003e, TOF (MIM: 187500), represented by the central node in the figure. TOF is a congenital heart defect characterised by a ventricular septal defect, pulmonary valve stenosis, thickened right ventricle and overriding aorta\u0026nbsp;(9). Patients with TOF develop cyanosis in proportion to the pulmonary valve stenosis, rapid breathing to compensate low oxygen levels and a heart murmur. Let us analyse each disease that we find connected to TOF in our similarity landscape. The \u003cem\u003eConotruncal Heart Malformations\u003c/em\u003e CHTM (MIM: 217095) disorder includes the TOF malformations and is known to be causally related to gene NKX2-5, a gene also known to be causally related to TOF. Both \u003cem\u003eAlagille Syndrome 1\u003c/em\u003e ALGS1 (MIM:118450) and \u003cem\u003eRight Atrial Isomerism\u003c/em\u003e RAI (MIM:208530) not only share phenotypic similarities with TOF such as pulmonary stenosis (ALGS1) and complete atrioventricular septal defects (RAI), but also have disease genes in common with TOF, namely JAG1 and GDF1\u0026nbsp;(10). \u003cem\u003eCongenital heart defects, Multiple Types\u003c/em\u003e CHTD6 (MIM: 613854) \u0026nbsp; (formerly \u003cem\u003eTransposition of the great arteries\u003c/em\u003e DTGA3) often have ventricular septal defects and associations between CHTD6 and the TOF-associated gene GDF1 have been reported in the literature\u0026nbsp;(11). \u003cem\u003eAortic Arch Interruption, Facial Palsy, Retinal Coloboma\u003c/em\u003e (MIM: 107550) exhibits symptomatic similarities with TOF, such as fatigue, rapid breathing, fast heart rate, low oxygen levels among others\u0026nbsp;(12). As is the case with TOF, the aortic arch interruption is characterised (among other features) by a ventricular septal defect. Finally, \u003cem\u003eTakayasu Arteritis\u003c/em\u003e (MIM: 207600) is an inflammatory disease of the arteries, with predilection for the aorta and its branches. The disease is characterised by lesions that can, among others, have stenotic qualities\u0026nbsp;(13).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eInterestingly, the diseases in the graph\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003ewithout a direct connection to TOF reflect not only their associations with their immediate neighbours but also, to some extent, with TOF. For example, \u003cem\u003eDiGeorge syndrome\u003c/em\u003e DGS (MIM: 188400) not only shares a gene with TOF (TBX1), but also the outflow tract defects present in DGS are associated with a higher incidence of conotruncal abnormalities\u0026nbsp;(14).\u003c/p\u003e\n\u003cp\u003eLanDis is a web application in which the user can interact with all the elements in the graph and the diseases can be repositioned either by dragging them or through several pre-defined layouts (circular, concentric, grid, breadth-first and force directed). Seamless exploration of the diseases similarity landscapes can be performed through the selection of any disease in the landscape. Every disease similarity landscape can be downloaded in publication-quality, high-resolution PNG images for offline analysis. Users can also select a disease and obtain a catalogue of those diseases most similar to it in a tabular format, as well as a detailed comparison between pairs of diseases \u0026ndash; \u003cstrong\u003eFigure 2\u003c/strong\u003e shows the \u003cem\u003eCompare\u003c/em\u003e page for TOF and ALGS1. For users who wish to use the Caniza similarity data as part of a larger pipeline, a CSV plain-text file is available from the download section of the website. To ease further exploration, LanDis links every MeSH term, disease and disease gene to its corresponding entry in the OMIM, UniProt and National Library of Medicine websites respectively.\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eLanDis offers a new perspective to explore disease similarity relationships. It is a simple and powerful tool which can be used for differential diagnosis as diseases that present similar molecular features will be assigned high similarity. Importantly, LanDis provides the user with a rationale for the results by making available the set of MeSH terms, corresponding to disease phenotypes, used to calculate the disease similarity. In this way, scientists can focus on the clinical features deemed more critical while concentrating on a selected list of highly similar diseases.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNotably, LanDis is able to find similarities at molecular level between diseases even in the absence of any molecular information \u0026ndash; this is because it only needs a list of publications associated with each disease. \u003cstrong\u003eSupplementary Figure 1\u0026nbsp;\u003c/strong\u003eshows the number of publications, MeSH terms and genes associated to the diseases in LanDis. As is expected, a disease with many referenced publications tends to be annotated by many MeSH terms, but a high number of publications does not necessarily correspond to a high number of known genes \u0026ndash; for example, Huntington\u0026rsquo;s disease, that has more than 450 references and close to a 1000 MeSH terms, is associated to a single gene. However, since LanDis relies exclusively on publications and their corresponding MeSH terms, the sparseness of molecular information does not prevent the similarity scores from being calculated. In fact, LanDis attempts to encapsulate all available information about diseases \u0026ndash; for example, the references of type 2 Diabetes (NIDDM) include information about several clinical trials and multi-year studies on the effects of glucose on insulin levels.\u003c/p\u003e\n\u003cp\u003eLanDis aims at becoming a support tool for bioinformaticians as well as medical practitioners. It is freely available through its website, no registration or installation is needed and our servers store no information about the users.\u003c/p\u003e"},{"header":"ONLINE METHODS","content":"\u003cp\u003e\u003cstrong\u003eDisease similarities and datasets\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLanDis mines OMIM to extract 139,549 PubMed references. For each publication, LanDis queries the Medline API obtaining a total of 17,110 MeSH terms. A few disease entries in OMIM with no references or MeSH annotations are excluded from LanDis, for a working total of 9,735 diseases. This amounts to over 44.7 million similarities, one per disease pair.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo produce the pairwise similarities, LanDis relies on the structure of the MeSH ontologies. The similarity between a pair of diseases is given by the Resnik similarity of the sets of MeSH terms annotating the diseases (16). The Resnik similarity score of two sets of MeSH terms is given by the information content of their lowest common ancestor, which is defined as the negative logarithm of the probability of finding it among the annotations of the OMIM diseases (16-18).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMeSH terms are organised into 16 ontologies and a given disease can be annotated with terms from more than one ontology. This means that for every disease up to 16 similarities can be calculated. Following Caniza \u003cem\u003eet al.\u003c/em\u003e (8), LanDis exploits the fact that these ontologies are interconnected to combine them and produce a single score.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eImplementation details\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLanDis is implemented using Python and the Django framework, following a strict Model-View-Controller architecture. The data persistence is provided by a single-file SQLite database, which holds the similarity data and all additional information required to provide LanDis functionalities. Indices where defined to improve access time to the SQL database. The user interface was designed using HTML 5 and the JQuery JavaScript library. Additionally, two well-known JavaScript libraries, D3.js and Cytoscape.js, are included. D3.js provides the tools for dynamic visualisations of the similarity data and Cytoscape.js provides the engine for LanDis disease landscape explorer. This allows for a flexible interface that fits most resolutions for desktops, laptops and most mobile devices.\u003c/p\u003e\n\u003cp\u003eThere are no special requirements for a user\u0026apos;s computer, since all user-side JavaScript code was carefully developed to reduce its footprint. Warnings are displayed for larger more resource-consuming plots, allowing the user to choose whether to continue with the operation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe source code is freely available from GitHub at https://github.com/paccanarolab/landis and is released under the GPLv3 license. We have tested LanDis on all major browsers and operating systems (mobile and desktop), and it performs best on Google Chrome.\u003c/p\u003e"},{"header":"DECLARATIONS","content":"\u003cp\u003eDr. Caniza has nothing to disclose.\u003c/p\u003e\n\u003cp\u003eDr. C\u0026aacute;ceres has nothing to disclose.\u003c/p\u003e\n\u003cp\u003eDr. Torres has nothing to disclose.\u003c/p\u003e\n\u003cp\u003eDr. Paccanaro has nothing to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003eFUNDING:\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA.P. was supported by Biotechnology and Biological Sciences Research Council (https://bbsrc.ukri.org/) grant numbers BB/K004131/1, BB/F00964X/1,and BB/M025047/1; Medical Research Council (https://mrc.ukri.org) grant number MR/T001070/1; Consejo Nacional de Ciencia y Tecnolog\u0026iacute;a Paraguay (https://www.conacyt.gov.py/) grants numbers 14-INV-088 (to AP, JC, MT and HC), PINV15\u0026ndash;315, and PINV20-337; National Science Foundation Advances in Bio Informatics (https://www.nsf.gov/) grant number 1660648; Fundac\u0026atilde;o de Amparo a Pesquisa do Estado do Rio de Janeiro (https://www.faperj.br) grant number E-26/201.079/2021 (260380); Conselho Nacional de Desenvolvimento Cient\u0026iacute;fico e Tecnol\u0026oacute;gico (https://www.cnpq.br) grant number 311181/2022-8; and Fundac\u0026atilde;o Getulio Vargas.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHC and AP developed the model. HC conceptualised the software. HC, JC and MT developed LanDis. AP tested and provided feedback on the features of LanDis. HC and AP wrote the manuscript. \u0026nbsp;All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Diego Galeano for useful discussions on the user interface.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBarab\u0026aacute;si A-L, Gulbahce N, Loscalzo J. Network medicine: a network-based approach to human disease. Nature Reviews Genetics. 2011;12(1):56-68.\u003c/li\u003e\n\u003cli\u003eXiujuan Wang NGaHY. Network-based methods for human disease gene prediction Briefings in Functional Genomics. 2011:280\u0026ndash;93.\u003c/li\u003e\n\u003cli\u003eZou Q, Li J, Wang C, Zeng X. Approaches for recognizing disease genes based on network. BioMed research international. 2014;2014.\u003c/li\u003e\n\u003cli\u003eZou Q, Li J, Song L, Zeng X, Wang G. Similarity computation strategies in the microRNA-disease network: a survey. Briefings in functional genomics. 2015;15(1):55-64.\u003c/li\u003e\n\u003cli\u003eZou Q, Li J, Hong Q, Lin Z, Wu Y, Shi H, et al. Prediction of microRNA-disease associations based on social network analysis methods. BioMed research international. 2015;2015. \u003c/li\u003e\n\u003cli\u003eGliozzo, J., Perlasca, P., Mesiti, M., Casiraghi, E., Vallacchi, V., Vergani, E., ... \u0026amp; Valentini, G. (2020). Network modeling of patients\u0026apos; biomolecular profiles for clinical phenotype/outcome prediction. Scientific Reports, 10(1), 3612.\u003c/li\u003e\n\u003cli\u003eC\u0026aacute;ceres, J. J., \u0026amp; Paccanaro, A. (2019). Disease gene prediction for molecularly uncharacterized diseases. PLoS computational biology, 15(7), e1007078.\u003c/li\u003e\n\u003cli\u003eCaniza H, Romero AE, Paccanaro A. A network medicine approach to quantify distance between hereditary disease modules on the interactome. Scientific reports. 2015;5.\u003c/li\u003e\n\u003cli\u003eOMIM. OMIM Entry 187500. Online Mendelian Inheritance in Man2017.\u003c/li\u003e\n\u003cli\u003eGruber PJ, Epstein JA. Development Gone Awry. Circulation Research. 2004;94:273-83.\u003c/li\u003e\n\u003cli\u003eKarkera JD, Lee JS, Roessler E, Banerjee-Basu S, Ouspenskaia MV, Mez J, et al. Loss-of-Function Mutations in Growth Differentiation Factor-1 (GDF1) Are Associated with Congenital Heart Defects in Humans. American journal of human genetics. 2007:81 (5): 987-94.\u003c/li\u003e\n\u003cli\u003eCollins-Nakai RL, Dick, M., Parisi-Buckley, L., Fyler, D. C., \u0026amp; Castaneda, A. R. Interrupted aortic arch in infancy. The Journal of pediatrics. 1976:88(6), 959-62.\u003c/li\u003e\n\u003cli\u003eSaruhan-Direskeneli G, Hughes, T., Aksu, K., Keser, G., Coit, P., Aydin, S. Z., ... \u0026amp; Hoffman, G. S. Identification of multiple genetic susceptibility loci in Takayasu arteritis. . The American Journal of Human Genetics,. 2013: 93(2), 298-305.\u003c/li\u003e\n\u003cli\u003eBruneau BG. The developmental genetics of congenital heart disease. Nature. 2008:451(7181), 943.\u003c/li\u003e\n\u003cli\u003eMcCright B, Lozier, J., \u0026amp; Gridley, T. A mouse model of Alagille syndrome: Notch2 as a genetic modifier of Jag1 haploinsufficiency. Development. 2002:129(4), 1075-82.\u003c/li\u003e\n\u003cli\u003eResnik, P. Semantic Similarity in a Taxonomy: An Information-Based Measure and its Applications to Problems of Ambiguity in Natural Language. Journal of Artificial Intelligence Research 11 (1999) 95-130.\u003c/li\u003e\n\u003cli\u003eYang H., Nepusz T., Paccanaro A. Improving GO semantic similarity measures by exploring the ontology beneath the terms and modelling uncertainty, Bioinformatics, Volume 28, Issue 10, May 2012, Pages 1383\u0026ndash;1389.\u003c/li\u003e\n\u003cli\u003e18. Caniza, H., Romero, A. E., Heron, S., Yang, H., Devoto, A., Frasca, M., ... \u0026amp; Paccanaro, A. (2014). GOssTo: a stand-alone application and a web tool for calculating semantic similarities on the Gene Ontology. Bioinformatics, 30(15), 2235-2236.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"european-journal-of-human-genetics","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"ejhg","sideBox":"Learn more about [European Journal of Human Genetics](http://www.nature.com/ejhg/)","snPcode":"41431","submissionUrl":"https://mts-ejhg.nature.com/cgi-bin/main.plex","title":"European Journal of Human Genetics","twitterHandle":"@ejhg_journal","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Disease similarity, genetic diseases, systems medicine, network medicine, differential diagnosis","lastPublishedDoi":"10.21203/rs.3.rs-3168447/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3168447/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eFrom a network medicine perspective, a disease is the consequence of perturbations on the interactome. These perturbations tend to appear in a specific neighbourhood on the interactome, the disease module, and modules related to phenotypically similar diseases tend to be located in close-by regions.\u003c/p\u003e\n\u003cp\u003eWe present LanDis, a freely available web-based interactive tool (paccanarolab.org/landis) that allows domain experts, medical doctors and the larger scientific community to graphically navigate the interactome distances between the modules of over 44 million pairs of heritable diseases. The map-like interface provides detailed comparisons between pairs of diseases together with supporting evidence. Every disease in LanDis is linked to relevant entries in OMIM and UniProt, providing a starting point for in-depth analysis and an opportunity for novel insight into the aetiology of diseases as well as differential diagnosis.\u003c/p\u003e","manuscriptTitle":"LanDis: The Disease Landscape Explorer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2023-08-28 14:11:22","doi":"10.21203/rs.3.rs-3168447/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"revise","date":"2023-10-03T12:58:40+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"This content is not available.","date":"2023-09-01T15:30:24+00:00","index":2,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2023-09-01T14:30:24+00:00","index":2,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2023-08-23T07:42:28+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2023-08-23T06:42:28+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewersInvited","content":"","date":"2023-08-22T09:25:58+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2023-07-24T08:45:42+00:00","index":"","fulltext":""},{"type":"submitted","content":"European Journal of Human Genetics","date":"2023-07-21T00:39:22+00:00","index":"","fulltext":""},{"type":"checksFailed","content":"","date":"2023-07-17T17:25:59+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2023-07-13T21:08:14+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"european-journal-of-human-genetics","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"ejhg","sideBox":"Learn more about [European Journal of Human Genetics](http://www.nature.com/ejhg/)","snPcode":"41431","submissionUrl":"https://mts-ejhg.nature.com/cgi-bin/main.plex","title":"European Journal of Human Genetics","twitterHandle":"@ejhg_journal","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"81067e99-1018-4367-9073-1e16ce996729","owner":[],"postedDate":"August 28th, 2023","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":24225696,"name":"Biological sciences/Computational biology and bioinformatics/Genome informatics"},{"id":24225697,"name":"Health sciences/Medical research/Genetics research"},{"id":24225698,"name":"Biological sciences/Computational biology and bioinformatics/Data processing"}],"tags":[],"updatedAt":"2024-01-22T18:58:11+00:00","versionOfRecord":{"articleIdentity":"rs-3168447","link":"https://doi.org/10.1038/s41431-023-01511-9","journal":{"identity":"european-journal-of-human-genetics","isVorOnly":false,"title":"European Journal of Human Genetics"},"publishedOn":"2024-01-10 05:00:00","publishedOnDateReadable":"January 10th, 2024"},"versionCreatedAt":"2023-08-28 14:11:22","video":"","vorDoi":"10.1038/s41431-023-01511-9","vorDoiUrl":"https://doi.org/10.1038/s41431-023-01511-9","workflowStages":[]},"version":"v1","identity":"rs-3168447","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3168447","identity":"rs-3168447","version":["v1"]},"buildId":"_2-kVJe1T_tPrBINL-cwx","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}