2 results
Genetic underpinnings of YMRS and MADRS scores variations in a bipolar sample
- M. Calabrò, C. Crisafulli, A. Drago
-
- Journal:
- European Psychiatry / Volume 66 / Issue S1 / March 2023
- Published online by Cambridge University Press:
- 19 July 2023, pp. S386-S387
-
- Article
-
- You have access Access
- Open access
- Export citation
-
Introduction
Bipolar disorder (BD) is a chronic hereditary disorder. Trial and error principles and long period of untreated disorder mandate further research. Relatively recent advances in statistical computing and techniques introduced Polygenic risk scores (PRS) as predictors of the genetic susceptibility to diseases. Although they provide an estimate of the risk of developing specific pathologies, they are a genome-wide measure. PRS do not provide specific information on the biological meaning of the variants. The use of subsets of risk variants (limited to one or few related biological pathways) to calculate pathway-PRS (pPRS) may provide an estimate of the functioning of specific molecular cascades.
ObjectivesIn the present study we calculated pPRS and tested them as potential predictive factors which, together with other clinical/environmental features, may estimate the treatment outcome of BD individuals in a clinical realistic treatment environment.
Methods1538 BD (41.39+/-12.66 years, 59.17% females) individuals from STEP-BD were included in the analysis. A latent class analysis identified three groups of patients according to the YMRS and MADRS scores variations during ˜ 1 year (308.47+/-293.83 days YMRS, 357.78+/-367.76 days MADRS). A GWAS analysis with clinical covariates provided the input for pPRS calculation. SNPs with best nominal significance and biologic relevance were prioritized through GTEx. A molecular pathway analysis (MPA) based on the interaction network of drugs used for treatment provided the genetic data needed for pPRS calculation. A Neural network was built using pPRS as features together with other variables (including Sex, Age, Scores at baseline) to predict the 3 groups previously identified. Performance was evaluated through 5-fold cross-validation, Python, R and Bash served for environments. Gene Ontology, ReactomePA and Bioconductor were key packages together with Cytoscape, Plink, impute and gtool.
ResultsTen biological networks were retrieved from MPA: 1)GO:0016705 + GO:0016641, 2)GO:0019585, 3)GO:003018, 4)GO:0099589 + GO:1904014, 5)GO:0015464 + GO:1905144, 6)GO:0004935 + GO:0004364 + GO:00031690, 7)GO:1903351 + GO:1903350, 8)GO:0016917 + GO:0007214, 9)GO:0008066 + GO:0007215, 10)GO:0048016. Risk variants within the genes contained in each group were used to compute pPRS. The ten pPRS were used to compute a neural network to predict treatment outcomes.
ConclusionsBD treatment is influenced by socio-demographic, clinical and genetic factors. To tackle this complexity, we tried to implement an approach where the multivariate analysis encompasses clinical analysis and the biologic background of treatment response. As a result, we can infer through a hypothesis-free approach potential pathways whose alterations may estimate treatment. At the time of writing the analyses are still undergoing, the final results will be presented and discussed at the congress.
Disclosure of InterestNone Declared
Inflammation and Pruning May Inform Risk to Psychiatric Disorders. Lessons From Large Genetic Data
- C. Crisafulli
-
- Journal:
- European Psychiatry / Volume 41 / Issue S1 / April 2017
- Published online by Cambridge University Press:
- 23 March 2020, p. S56
-
- Article
-
- You have access Access
- Export citation
-
Background
It's known that psychiatric disorders are caused to either environmental and genetics factors. Through the years several hypotheses were tested and many genes were screened for association, resulting in a huge amount of data available for the scientific community. Despite that, the molecular mechanics behind psychiatric disorders remains largely unknown. Traditional association studies may be not enough to pinpoint the molecular underpinnings of psychiatric disorder. We tried to applying a methodology that investigates molecular-pathway-analysis that takes into account several genes per time, clustered in consistent molecular groups and may successfully capture the signal of a number of genetic variations with a small single effect on the disease. This approach might reveal more of the molecular basis of psychiatric disorders.
Methodsi)We collected data on studies available in literature for the studied disorder (e.g. Schizophrenia, Bipolar Disorder);ii)We extracted a pool of genes that are likely involved with the disease;iii)We used these genes as starting point to map molecular cascades function-linked. The molecular cascades are then analyzed and pathways and sub-pathways, possibly involved with them, are identified and tested for association.
Results/discussionWe obtained interesting results. In particular, signals of enrichment (association) were obtained multiple times on the molecular pathway associated with the pruning activity and inflammation. Molecular mechanics related to neuronal pruning were focused as a major and new hypothesis for the pathophysiology of psychiatric disorders and the role of inflammatory events has been extensively investigated in psychiatry. intersting, inflammatory mechanics in the brain may also play a role in neuronal pruning during the early development of CNS.
Disclosure of interestThe author has not supplied his declaration of competing interest.