From de4565fe9290ec1f1031eed6f7d067794df53166 Mon Sep 17 00:00:00 2001
From: Mike Vink <mike1994vink@gmail.com>
Date: Wed, 28 Apr 2021 18:21:50 +0200
Subject: begin data prep

---
 data_preparation/main.acn |   0
 data_preparation/main.pdf | Bin 0 -> 520861 bytes
 data_preparation/main.tex |  53 ++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 53 insertions(+)
 create mode 100644 data_preparation/main.acn
 create mode 100644 data_preparation/main.pdf

(limited to 'data_preparation')

diff --git a/data_preparation/main.acn b/data_preparation/main.acn
new file mode 100644
index 0000000..e69de29
diff --git a/data_preparation/main.pdf b/data_preparation/main.pdf
new file mode 100644
index 0000000..f401b33
Binary files /dev/null and b/data_preparation/main.pdf differ
diff --git a/data_preparation/main.tex b/data_preparation/main.tex
index 9c729b3..5186dc3 100644
--- a/data_preparation/main.tex
+++ b/data_preparation/main.tex
@@ -15,4 +15,57 @@
 
 \section{Data selection}
 
+The data that we use in this work is based on the data used in the simon
+manuscript \citep{tomicSIMONAutomatedMachine2019}. This subset of the fluprint
+database comprises data from 5 clinical studies, most importantly the
+longitudal study SLVP015. It only uses the first visits of donors, as the
+classification is the most complete in this dataset
+\autoref{fig:dataRepeatvisits}. We will use this dataset to model the high
+vaccine response with as predictors the in total 3285 features measured in
+assays done by the clinical studies. The data will be prepared in the same way
+as in the original work, using the mulset algorithm
+\citep{tomicSIMONAutomatedMachine2019}.
 
+\begin{figure}
+    \includegraphics[width=\textwidth]{data_selection}
+    \caption{caption}\label{fig:dataRepeatvisits}
+\end{figure}
+
+In addition to repeating a similar procedure as in
+\cite{tomicSIMONAutomatedMachine2019}, we will compare the values of features
+selected by the models trained on the first visit data to those of second visit
+data. Initially the plan was to train new models on the second visit data,
+however the classes are extremely unbalanced in the data of repeat visits
+\autoref{fig:dataRepeatvisits}. For example in the first visit there are 65
+high responders and 130 low responders, in the second visit there is only data
+available for 6 high responders and 44 low responders. Therefore we will only
+train model on first visit data, and use the knowledge gained to explore second
+visit data.
+
+\section{Clean data}
+
+The goal is to obtain one or more tables from the simon data suitable to train
+models, thus we are looking to change the data from the long format as in the
+database in a wide format where each column is a features measured in an assay.
+When attempting to do this it was discovered that some assay data contained
+duplicate readouts \autoref{tbl:exampleDuplicate}. Since the values were all
+similar it was decided to aggregate the values to unique features using the
+mean value.
+
+\begin{table}
+\addtolength{\leftskip} {-2cm} % increase (absolute) value if needed
+\addtolength{\rightskip} {-2cm} % increase (absolute) value if needed
+\begin{tabular}{rrrrrlrllrrl}
+\toprule{}
+donor\_id & study & age & outcome & year & type & hai\_response & name & data\_name & assay & data & dup\\
+\midrule{}
+285 & 18 & 9.47 & 0 & 2009 & pre & 1 & CD4+ T cells & CD4\_pos\_T\_cells & 13 & 33.8 & TRUE\\
+285 & 18 & 9.47 & 0 & 2009 & pre & 1 & CD4+ T cells & CD4\_pos\_T\_cells & 13 & 34.1 & TRUE\\
+285 & 18 & 9.47 & 0 & 2009 & pre & 1 & CD4+ T cells & CD4\_pos\_T\_cells & 13 & 34.3 & TRUE\\
+285 & 18 & 9.47 & 0 & 2009 & pre & 1 & CD4+ T cells & CD4\_pos\_T\_cells & 13 & 33.0 & TRUE\\
+\bottomrule{}
+\end{tabular}
+    \caption{}\label{tbl:exampleDuplicate}
+\end{table}
+
+\end{document}
-- 
cgit v1.2.3