evotec_Nov23_PanOmicsDriven - 39

research, PanHunter allows to
search for driving factors of
sample distribution: Exploratory
Analysis. It can be used to search
for parameters in the metadata
of the samples, or even in the
expression values of the genes
themselves, that are responsible
for the distribution of the
samples across the 2D plot,
using statistical methods. As a
result, a list of possible genes
was generated, showing REG4 on
top (Figure 8).
To examine this result in a
bit more detail, the coloration
of the sample plot was changed
to indicate the transcriptional
intensity of that gene (Figure 9).
The resulting plot showed a
clear gradient within the large
cloud of samples, from low (light
green) to high (dark green) transcriptional
activity of REG4. The
correlation of REG4 expression
with the overall clustering of the
patients was a first hint that the
gene could be a potential marker
gene.
A comparison between the
activity of REG4 and survival of
patients was performed in the
Patient Data app: First, the patient
cohort was divided into two
sub-cohorts via a respective selection
in the graphical user interface
(GUI). Subsequently, one sub-coPval
log-rank: 0.0114; Pval CoxPH Wald: 0.0711
hort containing the lower 50%
(red) and another one containing
the upper 50% quantile (cyan)
with respect to their REG4 expression
were generated. As a result,
a Kaplan-Meier plot was generated,
indicating patient survival
for both sub-cohorts. In the plot,
a slight difference between the
sub-cohorts was already visible:
patients with high REG4 expresREG4
(ENSG00000134193)
Figure 10: Kaplan-Meier-Plot for male patients. Here, we see a significant
(p-value = 0.0114) correlation between low REG4 transcription (red curve) and poor survival.
39
GENengnews.com
Survival

https://www.evotec.com/en https://www.genengnews.com https://www.genengnews.com

evotec_Nov23_PanOmicsDriven

Table of Contents for the Digital Edition of evotec_Nov23_PanOmicsDriven