A dynamic network approach for the study of human phenotypes.
Hidalgo CA, Hidalgo CA, Blumm N, Barabási AL, Christakis NA. · 2009
Read A dynamic network approach for the study of human phenotypes. by Hidalgo CA, Hidalgo CA, Blumm N, Barabási AL, Christakis NA. (2009) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc2661364.txt.
Introduction
Hidalgo and colleagues built in 2009 a network of human diseases from the medical histories of more than 30 million patients. Diseases that often strike the same people are linked, forming a map of how illnesses cluster and follow one another. The team released the map, called the Phenotypic Disease Network, as a public resource.
Clear patterns emerged from the map. Patients tended to develop new diseases close to ones they already had. The paths of spread differed between men and women, and between ethnic groups. People whose diseases sat at busy hubs of the map tended to die sooner than those with lonely, rarely linked diseases. Diseases that usually came late in a chain were also more connected and more deadly. Disease, the authors argue, advances along this map like traffic along roads.
Catalog record
| CREATOR | Hidalgo CA, Hidalgo CA, Blumm N, Barabási AL, Christakis NA. |
|---|---|
| YEAR | 2009 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 41 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A Dynamic Network Approach for the Study of Human Phenotypes By César A. Hidalgo, Nicholas Blumm, Albert-László Barabási, Nicholas A. Christakis, Lauren Ancel Meyers Published in: PLoS Computational Biology DOI: 10.1371/journal.pcbi.1000353 ABSTRACT The use of networks to integrate different genetic, proteomic, and metabolic datasets has been proposed as a viable path toward elucidating the origins of specific diseases. Here we introduce a new phenotypic database summarizing correlations obtained from the disease history of more than 30 million patients in a Phenotypic Disease Network (PDN). We present evidence that the structure of the PDN is relevant to the understanding of illness progression by showing that (1) patients develop diseases close in the network to those they already have; (2) the progression of disease along the links of the network is different for patients of different genders and ethnicities; (3) patients diagnosed with diseases which are more highly connected in the PDN tend to die sooner than those affected by less connected diseases; and (4) diseases that tend to be preceded by others in the PDN tend to be more connected than diseases that precede other illnesses, and are associated with higher degrees of mortality. Our findings show that disease progression can be represented and studied using network methods, offering the potential to enhance our understanding of the origin and evolution of human diseases. The dataset introduced here, released concurrently with this publication, represents the largest relational phenotypic resource publicly available to the research community. INTRODUCTION There are no clear boundaries between many diseases, as diseases can have multiple causes and can be related through several dimensions. From a genetic perspective, a pair of diseases can be related because they have both been associated with the same gene [1],[2], whereas from a proteomic perspective diseases can be related because disease associated proteins act on the same pathway [3]–[9]. During the past half-decade, several resources have been constructed to help understand the entangled origins of many diseases. Many of these resources have been presented as networks in which interactions between disease-associated genes, proteins, and expression patterns have been summarized. For example, Goh et al. created a network of Mendelian gene-disease associations by connecting diseases that have been associated with the same genes [1] (see also Feldman et al. [2]), whereas Lee et al. constructed a network in which two diseases are linked if mutated enzymes associated with them catalyze adjacent metabolic reactions [10]. Network studies in the proteomic front have studied large protein interaction networks, like the ones created by Rual et al. [3] and Stelzl et al. [4], in an attempt to understand diseases like inherited ataxias [5] or Huntington's disease [6]. Moreover, in the gene expression front, microarray expression profiles and other cellular level information have been used to explore networks in inflammation [7], breast cancer [8], and brain disease [9]. While progress on the genetic and proteomic fronts has been impressive [1],[10], much of the available resources overlook the fact that we have extensive and continually updated phenotypic information for humans – namely, patient clinical histories. Indeed, hospitals and insurance programs constantly collect detailed records for millions of patients. These datasets contain information on disease associations and progression. For example, such population-based disease associations could be used in conjunction with molecular and genetic data to help us uncover the molecular origins of diseases. Despite the potential utility of population based disease associations, extensive datasets linking diseases based on comorbidity associations do not exist, partly because access to extensive medical records is limited. Typically, we say that a comorbidity relationship exists between two diseases whenever they affect the same individual substantially more than chance alone. One of our primary goals here is to make available pairwise comorbidity correlations for more than 10 thousand diseases reconstructed from over 30 million medical records. For completeness and utility, we organize the results in 18 different datasets. Each summarizes phenotypic associations extracted from four years worth of ICD9-CM claims data at the 5 and 3 digit level. Results are grouped into subsets of race, gender, and both race and gender (see SM). To facilitate their use, the datasets are available as a bulk download (http://hudine.neu.edu/resource/data/data.html) or through a searchable web interface (http://hudine.neu.edu) that allows researchers, doctors and patients to explore these disease networks graphically, through an interactive Flash application, and numerically, by allowing them to generate tables summarizing the associations between a particular disease and all other diseases. In the past, comorbidities have been used extensively to construct synthetic scales for mortality prediction [11],[12], yet their utility could exceed their current use. Studying the structure defined by entire sets of comorbidities might help the understanding of many biological and medical questions from a perspective that is complementary to other approaches. For example, a recent study built a comorbidity network in an attempt to elucidate neurological diseases common genetic origins [13]. Heretofore, however, neither this data nor the data necessary to explore relationships between all diseases is currently available to the research community. Hence, here we decide to provide this data in the form of a Phenotypic Disease Network (PDN) capturing all diseases as recorded through medical claims. Additionally, we illustrate how a PDN can be used to study illness progression from a dynamic network perspective by interpreting the PDN as the landscape where illness progression occurs and show how the network can be used to study phenotypic differences between patients with different demographic backgrounds. Furthermore, we show that the local structure of a disease in the network, as characterized by its degree or number of connections, is associated with disease mortality. Finally, we study the directionality of disease progression, as observed in our dataset, and find that more central diseases in the PDN are more likely to occur after other diseases and that more peripheral diseases tend to precede other illnesses. We also find that patients diagnosed with diseases that tend to be preceded by other conditions tend to die sooner than those diagnosed with conditions that tend to precede other diseases. Together, these results and resources open new opportunities for biomolecular, bioinformatic and public health approaches to disease. METHODS Source Data and Study Population Hospital claims offer reliable, systematic, and complete data for disease detection [14]–[16]. Each record consists of the date of visit, a primary diagnosis and up to 9 secondary diagnoses, all specified by ICD9 codes of up to 5 digits. The first three digits specify the main disease category while the last two provide additional information about the disease. In total, the ICD-9-CM classification consists of 657 different categories at the 3 digit level and 16,459 categories at 5 digits. For a detailed list of currently used ICD9 codes see www.icd9data.com. We compiled raw Medicare claims [17],[18] based on so-called MedPAR records regarding hospitalizations for 1990–1993. Medicare is the US government's health insurer, and it has information on 96% of all elderly Americans whether they seek medical care or not [19]. For the 32 million elderly Americans aged 65 or older enrolled in Medicare and alive for the entire study period, there were a total of 32,341,347 inpatient claims, pertaining to 13,039,018 individuals (the remaining individuals were not hospitalized at any point during this period). Demographically, our data set consists of