| 摘要 |
#CMT# #/CMT# Computer-assisted analysis of an interaction network of biomedical entities, where: e.g. the biomedical entities are represented by nodes and the biomedical entities comprise many first entities and one or more second entities; and at least one second entity represents a disease in which for many combinations of first entities the following steps are carried out by: (a) modeling interventions on conditions of the first entities of the respective combinations and (b) determining the influence of the interventions on the conditions of the one or more second entities, is claimed. #CMT# : #/CMT# Computer-assisted analysis of an interaction network of biomedical entities, where: the biomedical entities are represented by nodes and the influences between the biomedical entities are represented by edges in the interaction network and the biomedical entities comprise many first entities and one or more second entities; preferably one or more of the first entities represent biomolecules; and at least one second entity represents a disease in which for many combinations of first entities the following steps are carried out by: (a) modeling interventions on conditions of the first entities of the respective combinations and (b) determining the influence of the interventions on the conditions of the one or more second entities and characterizing the respective combinations when the influence of the interventions exceeds a predetermined dimension, is claimed. An independent claim is included for a computer program product with a program code stored on a machine-readable carrier for carrying out the above method when the program runs in a processor. #CMT#USE : #/CMT# The method is useful for the computer-assisted analysis of an interaction network of biomedical entities. #CMT#ADVANTAGE : #/CMT# The method analyzes a large number of different biomedical entities with regard to its impact on a disease. #CMT#BIOLOGY : #/CMT# Preferred Components: The first entities represent one or more of the following values: proteins, genes and/or metabolites; clinical data of patients, preferably blood corpuscles concentration, histological findings, laboratory findings and/or anamnesis and/or patient history; environmental factors, which influences a patient; and nutrition of the patient. The interaction network is based on a probability network, preferably on a Bayesian network, on expert knowledge and/or from the technical literature extracted information. The effect of the interventions on the conditions of one or more second entities is described by intervention scores. The conditions of the biomedical entities and the interaction between the biomedical entities are connected with statistical values, preferably probabilities, probability distributions, statistical correlation values, confidences and/or statistical values for the interaction of a node with other nodes. The conditions of the at least one disease representing second entity displays the occurrence or absence of the disease. The first entities comprise genes with predetermined gene expression conditions. The gene expression conditions comprise the conditions of normal expressed, over expressed and under expressed. Preferred Method: In step (b), the first entities of each combination and their conditions as well as influences among each other and/or with other biomedical entities of the interaction network are determined by a characteristic combination. The modeling of the interventions on the conditions of the first entities simulates a respective combination of a medicament administration on a patient. At least one second entity represents the side-effect or side-effects of the medicament administration. The modeling of the interventions on the conditions of the gene comprises the adjusting of a fixed gene expression conditions for the gene. The combinations of first biomedical entities and the interventions carried out on it are determined iteratively depending on the already used combinations and interventions. The iteration is carried out with the help of an optimization procedure, which allows the influence on the interventions on one or more second entities, which are determined in one or more previous iteration steps, and determines a new combination of first entities and interventions carried out on it. The optimization procedure is based on Greedy Hill Climbing and/or simulated annealing. In the method, all combinations of first entities are processed in the steps (a) and (b). The molecular paths are contained in the interaction network between the first entities and in step (a) the interventions are modeled, which simultaneously act upon the conditions of the first entity or entities of a single molecular path or many molecular paths. The interactions network is displayed on a display tool as graphical net structure, in which the biomedical entities are visualized as nodes and the interactions between the biomedical entities are visualized as edges between the nodes, where in the graphical net structure, the conditions of the biomedical entities and the values of the interactions are encoded between the biomedical entities. |
