Study name: r params$study
Array type: r params$array
Normalization method: r params$normalization
Nuber of CpGs not in the study: r params$numberCpG
Contact: r params$email
In this project, we aim to validate a machine learning based DNA-methylation exposure score to maternal smoking in utero. This score was developed in the Raine Study (partner: University of Western Australia/Telethon Kids Instiute, Perth, Western Australia) and the Northern Finland Birth Cohorts 1986 and 1966 (partner: University of Oulu, Oulu, Finland). For the score development, DNA methylation data measured with the Illumina HumanMethylation450 BeadChip array was utilized, together with a binary variable, indicating if the study participants mother was smoking during pregnancy.
The resulting score outperforms currently existing alternatives and adds a systematic evaluation of different machine learning approaches to the literature. In parallel to the development of the predictive score, a R-package called DNAsmokeR
was developed (https://github.com/Hobbeist/DNAsmokeR), that allows users to apply the method to their study and receive the score together with measures on the predictive power, given the study has information on maternal smoking during pregnancy available.
This is the validation report for: r params$study
The data that was used to create the predictive score needed to be without missing values. Therefore, NA values are excluded
in the modelling process.
r params$study
has r params$sampsize
subjects available without missing data. That is the number of participants for whom the score could be created.
knitr::opts_chunk$set(echo = TRUE)
The following confusion Matrix is created by the confusionMatrix()
function within the caret
package.
params$prediction
This is the ROC curve for the predictor:
params$rocCurve
r params$missingCpG
library(pander) pander(sessionInfo(), compact = FALSE)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.