Limits...
Enrichment analysis applied to disease prognosis.

Machado CM, Freitas AT, Couto FM - J Biomed Semantics (2013)

Bottom Line: With this analysis the objective is to identify clinical and biological features that characterize groups of patients with a common disease, and that can be used to distinguish between groups of patients associated with disease-related events.These analyses correspond to an adaptation of the standard enrichment analysis, since multiple sets of genes are being considered, one for each patient.The preliminary results are promising, as the sets of terms obtained reflect the current knowledge about the gene functions commonly altered in HCM patients, thus allowing their characterization.One of such factors is the need to test the enrichment analysis with clinical data, in addition to genetic data, since both types of data are expected to be necessary for prognosis purposes.

View Article: PubMed Central - HTML - PubMed

Affiliation: LaSIGE, Departamento de Informática, Faculdade de Ciências, Universidade de Lisboa, Lisboa, Portugal. cmachado@xldb.di.fc.ul.pt.

ABSTRACT
: Enrichment analysis is well established in the field of transcriptomics, where it is used to identify relevant biological features that characterize a set of genes obtained in an experiment.This article proposes the application of enrichment analysis as a first step in a disease prognosis methodology, in particular of diseases with a strong genetic component. With this analysis the objective is to identify clinical and biological features that characterize groups of patients with a common disease, and that can be used to distinguish between groups of patients associated with disease-related events. Data mining methodologies can then be used to exploit those features, and assist medical doctors in the evaluation of the patients in respect to their predisposition for a specific event.In this work the disease hypertrophic cardiomyopathy (HCM) is used as a case-study, as a first test to assess the feasibility of the application of an enrichment analysis to disease prognosis. To perform this assessment, two groups of patients have been considered: patients that have suffered a sudden cardiac death episode and patients that have not.The results presented were obtained with genetic data and the Gene Ontology, in two enrichment analyses: an enrichment profiling aiming at characterizing a group of patients (e.g. that suffered a disease-related event) based on their mutations; and a differential enrichment aiming at identifying differentiating features between a sub-group of patients and all the patients with the disease. These analyses correspond to an adaptation of the standard enrichment analysis, since multiple sets of genes are being considered, one for each patient.The preliminary results are promising, as the sets of terms obtained reflect the current knowledge about the gene functions commonly altered in HCM patients, thus allowing their characterization. Nevertheless, some factors need to be taken into consideration before the full potential of the enrichment analysis in the prognosis methodology can be evaluated. One of such factors is the need to test the enrichment analysis with clinical data, in addition to genetic data, since both types of data are expected to be necessary for prognosis purposes.

No MeSH data available.


Related in: MedlinePlus

Schematic representation of the prognosis methodology. The methodology is composed by two units: the first (left-side) receives as input data from patients mapped to biomedical ontologies/controlled vocabularies. It performs an enrichment analysis to identify a list of ontology terms considered to be enriched, which will be used to create profiles for individual patients. These profiles will then be subjected to an evaluation step (the second unit, on the right-side) that will result in the evaluation of the prognosis for the patients. For the implementation of the second unit, both a classification and a similarity approach will be explored.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4126066&req=5

Figure 1: Schematic representation of the prognosis methodology. The methodology is composed by two units: the first (left-side) receives as input data from patients mapped to biomedical ontologies/controlled vocabularies. It performs an enrichment analysis to identify a list of ontology terms considered to be enriched, which will be used to create profiles for individual patients. These profiles will then be subjected to an evaluation step (the second unit, on the right-side) that will result in the evaluation of the prognosis for the patients. For the implementation of the second unit, both a classification and a similarity approach will be explored.

Mentions: This article proposes the application of enrichment analysis for disease prognosis, as the first component of a prognosis methodology that will assist in the evaluation of patients in respect to the likelihood of suffering a disease-related event (see Figure 1). By performing an enrichment analysis on the patients’ data based on controlled vocabularies, we expect to identify sets of characterizing features that will be used as profiles for the patients. These profiles will then be explored to evaluate the predisposition of the patients for the specific event. This evaluation is the second component of the prognosis methodology, and can be performed by following a classification or a similarity approach. In the classification approach, the terms composing the profiles will be added as features to the patients’ dataset and analyzed with classification algorithms such as random forests [12] and Bayesian networks [13]. In the similarity approach, semantic similarity measures will be used to compute the similarity between patients, based on their profiles. Different semantic similarity measures [14] and a relatedness measure [15] can be explored to compare the patients’ profiles.


Enrichment analysis applied to disease prognosis.

Machado CM, Freitas AT, Couto FM - J Biomed Semantics (2013)

Schematic representation of the prognosis methodology. The methodology is composed by two units: the first (left-side) receives as input data from patients mapped to biomedical ontologies/controlled vocabularies. It performs an enrichment analysis to identify a list of ontology terms considered to be enriched, which will be used to create profiles for individual patients. These profiles will then be subjected to an evaluation step (the second unit, on the right-side) that will result in the evaluation of the prognosis for the patients. For the implementation of the second unit, both a classification and a similarity approach will be explored.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4126066&req=5

Figure 1: Schematic representation of the prognosis methodology. The methodology is composed by two units: the first (left-side) receives as input data from patients mapped to biomedical ontologies/controlled vocabularies. It performs an enrichment analysis to identify a list of ontology terms considered to be enriched, which will be used to create profiles for individual patients. These profiles will then be subjected to an evaluation step (the second unit, on the right-side) that will result in the evaluation of the prognosis for the patients. For the implementation of the second unit, both a classification and a similarity approach will be explored.
Mentions: This article proposes the application of enrichment analysis for disease prognosis, as the first component of a prognosis methodology that will assist in the evaluation of patients in respect to the likelihood of suffering a disease-related event (see Figure 1). By performing an enrichment analysis on the patients’ data based on controlled vocabularies, we expect to identify sets of characterizing features that will be used as profiles for the patients. These profiles will then be explored to evaluate the predisposition of the patients for the specific event. This evaluation is the second component of the prognosis methodology, and can be performed by following a classification or a similarity approach. In the classification approach, the terms composing the profiles will be added as features to the patients’ dataset and analyzed with classification algorithms such as random forests [12] and Bayesian networks [13]. In the similarity approach, semantic similarity measures will be used to compute the similarity between patients, based on their profiles. Different semantic similarity measures [14] and a relatedness measure [15] can be explored to compare the patients’ profiles.

Bottom Line: With this analysis the objective is to identify clinical and biological features that characterize groups of patients with a common disease, and that can be used to distinguish between groups of patients associated with disease-related events.These analyses correspond to an adaptation of the standard enrichment analysis, since multiple sets of genes are being considered, one for each patient.The preliminary results are promising, as the sets of terms obtained reflect the current knowledge about the gene functions commonly altered in HCM patients, thus allowing their characterization.One of such factors is the need to test the enrichment analysis with clinical data, in addition to genetic data, since both types of data are expected to be necessary for prognosis purposes.

View Article: PubMed Central - HTML - PubMed

Affiliation: LaSIGE, Departamento de Informática, Faculdade de Ciências, Universidade de Lisboa, Lisboa, Portugal. cmachado@xldb.di.fc.ul.pt.

ABSTRACT
: Enrichment analysis is well established in the field of transcriptomics, where it is used to identify relevant biological features that characterize a set of genes obtained in an experiment.This article proposes the application of enrichment analysis as a first step in a disease prognosis methodology, in particular of diseases with a strong genetic component. With this analysis the objective is to identify clinical and biological features that characterize groups of patients with a common disease, and that can be used to distinguish between groups of patients associated with disease-related events. Data mining methodologies can then be used to exploit those features, and assist medical doctors in the evaluation of the patients in respect to their predisposition for a specific event.In this work the disease hypertrophic cardiomyopathy (HCM) is used as a case-study, as a first test to assess the feasibility of the application of an enrichment analysis to disease prognosis. To perform this assessment, two groups of patients have been considered: patients that have suffered a sudden cardiac death episode and patients that have not.The results presented were obtained with genetic data and the Gene Ontology, in two enrichment analyses: an enrichment profiling aiming at characterizing a group of patients (e.g. that suffered a disease-related event) based on their mutations; and a differential enrichment aiming at identifying differentiating features between a sub-group of patients and all the patients with the disease. These analyses correspond to an adaptation of the standard enrichment analysis, since multiple sets of genes are being considered, one for each patient.The preliminary results are promising, as the sets of terms obtained reflect the current knowledge about the gene functions commonly altered in HCM patients, thus allowing their characterization. Nevertheless, some factors need to be taken into consideration before the full potential of the enrichment analysis in the prognosis methodology can be evaluated. One of such factors is the need to test the enrichment analysis with clinical data, in addition to genetic data, since both types of data are expected to be necessary for prognosis purposes.

No MeSH data available.


Related in: MedlinePlus