Clinical (liver cohort)¶
Single nucleotide ploymorphisms (SNPs) are single basepair mutations in DNA that occur with certain frequency within the population. These mutations might be the cause of a disease or due to their close vecinity to genetic abration they might be co-transfered with the cause.
In the study by Schadt EE et al., 2008 more than 780 thousand SNPs were analyzed in more than 400 human liver samples. Since a mutation could cause a phenotype that in time can cause a disease, the phenotypic information about the patients was included as well.
In this workflow, these data will be used as input for machine learning classification using SVM (from e1071 package). Data is devided into training and test sets and predictive models are generated for different features, these models are then tested using test dataset and the results are stored in a list.
![digraph CLINICAL_liver {
{rank=same train test}
{rank=same svm_gender_train svm_tg_train}
{rank=same pred_gender_train pred_gender_test pred_tg_train pred_tg_test}
data [shape = invhouse, label = "Liver cohort data \nread.delim()"];
sample [shape = box; label = "sample()"];
train [label = "Training set "];
test [label = "Test set "];
svm_gender_train [shape = box; label = "svm(y=gender) "];
pred_gender_train [shape = box; label = "predict() "];
pred_gender_test [shape = box; label = "predict() "];
agree_gender_train [shape = box; label = "classAgreement() "];
agree_gender_test [shape = box; label = "classAgreement() "];
svm_tg_train [shape = box; label = "svm(y=triglyceride) "];
pred_tg_train [shape = box; label = "predict() "];
pred_tg_test [shape = box; label = "predict() "];
rbind [shape = box; label = "do.call(rbind) "];
print [shape = box; label = "print() "];
data -> sample -> train -> svm_gender_train -> pred_gender_train -> agree_gender_train -> rbind -> print;
sample -> test -> pred_gender_test -> agree_gender_test;
train -> svm_tg_train -> pred_tg_train;
test -> pred_tg_test;
svm_gender_train -> pred_gender_test;
svm_tg_train -> pred_tg_test;
agree_gender_test -> rbind;
pred_tg_train -> rbind;
pred_tg_test -> rbind;
}](../../_images/graphviz-f20db368ed2471451aa89e615d24adcd368db537.png)
Analytic workflow for clinical liver cohort.
Packages and Dependencies¶
There is 1 package used in this workflow, which depends on 2 additional packages from CRAN (dependencies)
Used packages:
- CRAN: e1071
Package dependencies:
- CRAN: class, MASS
Data¶
Downloaded from https://www.synapse.org/ in June 2015
License¶
- Copyright (c) 2015 Ieuan Clay based on code from genbench
- Copyright (c) 2015-2016 BeDataDriven B.V. License: GPL version 2 or higher