Data Graphs for Linking Clinical Phenotype and Molecular Feature Space

Abstract

Omics profiling in translational clinical research has provided detailed molecular characterization of disease phenotypes. Integrating this molecular data space with clinical phenotype descriptors has triggered advancements regarding a systems view on disease, resulting in the concept of stratified medicine. The authors present a methodology for patient stratification by analyzing clinical and molecular information on a per-patient level represented as a data graph. This approach rests on linking patient specific clinical data and biomarker profiles with molecular functional units being derived by segmenting a human proteome interaction network. As a result patient strata are built holding sets of affected functional molecular units as common denominator. Annotation of such functional units on the level of associated diseases, biomarkers and drug targets allows reconciliation with respective clinical data for further improving the assignment of patients to specific strata. The authors finally discuss this approach in the light of adaptive clinical trials design and analysis.

Mass Spectrometric/Bioinformatic Identification of a Protein Subset That Characterizes the Cellular Activity of Anticancer Peptides

“…The interaction data from these three sources were consolidated in the omicsNET protein framework, which takes into account direct protein−protein and computationally calculated interactions based on shared pathway information, gene ontology annotations, and common protein domains. 21,22 In a second step, we extended the set of DEPs via an interneighbor expansion method, adding proteins to the initial set of DEPs that showed at least two interactions with functional neighbors through querying the BioGRID, IntAct, and Reactome databases. The rationale for this expansion was the goal of identifying proteins that were of relevance due to their strong connectivity to members of the DEP set that were too diluted or suppressed by more abundant ions to be identified via mass spectrometry.…”

Molecular models of the cardiorenal syndrome

“…On the other hand, the concept is versatile, allowing an integration of different Omics profiles for truly covering genetic predisposition as well as environmental impact in focus of risk for developing disease and disease progression . Ultimately, the true strength of molecular models may be the more complete representation of pathophysiology for given disease terminology as an assembly of causative processes which, on a population level, show patient‐specific contributions and relevance. By including the total set of processes, and consequently biomarker panels instead of single markers, present limitations in sensitivity and specificity in diagnosis and prognosis in complex disorders as the CRS may be overcome.…”

“…As recently demonstrated for DN , and laid out conceptually in Heinzel et al. , molecular models of disease terms, composed of a set of disease‐associated processes and their interrelation, promise improved interpretation of molecular mechanisms regarding their causality as well as relative contribution to a clinical presentation. Further, such molecular models may improve biomarker candidate selection, allowing monitoring the apparently complex, multi‐process pathophysiology of the CRS, in turn also allowing hypothesis generation regarding novel therapy targets.…”

From molecular signatures to predictive biomarkers: modeling disease pathophysiology and drug mechanism of action

“…Here, Omics profiling and high throughput drug screening technologies at the interface of large scale clinical data have triggered novel conceptual strategies aimed at improved patient stratification for enabling precision medicine (Trusheim et al, 2011 ; Hollebecque et al, 2014 ). For implementing such approaches a number of issues need to be addressed including: (i) mirroring the clinical categorization of a phenotype on a molecular level description, (ii) spotting molecular factors mechanistically driving disease progression, (iii) drug-based intervention specifically addressing such progression mechanisms, and (iv) predictive biomarkers allowing fit-for-purpose analysis regarding a match of relevant pathophysiology and drug mechanism of action on the individual patient level (Heinzel et al, 2012 ).…”