The protein subcellular location image database (http://pslid.
org/), The human protein data base atlas (http://pro
teinaltalas.org), LOCATE (http://locate.imb.uq.edu.au/).
Besides, there are many integrative platforms for the image
data management. For example OMERO or BISQUE or
KNIME is used for bioimage analysis. Interestingly, machine
learning powerful microscopy analysis is concerned with pow-
erful machine learning algorithms, which automatically classify
the pixels of the images and applied to the experiments for
identifying the microscopy images [24]. In the last two dec-
ades, several groups have innovated various collections of sci-
entific images for biological experiments. Workflow systems
tools have also enabled serving multiple data resources. Open
source workflow tools are used as Taverna (http://www.
taverna.org.uk/)and Galaxy (http://galaxy.psu.edu/).- Let us start with one confocal image of the sample from our
experimental data. First, we load the data into ImageJ. For this
purpose, we use ImageJ built-it functions: File --- open ---
image from folder. Further, we split the channels by using the
functions inside the ImageJ session through, file --- image ---
Color --- Split channels (Fig.7).
It is important to split the channels to separate the nuclei
and microtubules and to extract the resulting information.
Next, we contour the nuclei and microtubules to measure the
cell parameter that we choose to characterize morphological
and phenotypic features (Fig.8).
We can also use split images from the confocal data to
overlay the nuclei over the underlying microtubules (Fig.9).
Second, we analyze the nuclei that we have loaded before
and saved into the folder with the nuclei Region Of Interest
(ROI) manager tool. Once the threshold parameters have been
accomplished, it is of interest to analyze the effects and the
different structural dynamics onto the biological processes in
which the evaluations of the model were hypothesized. We
have adjusted the nuclei threshold at 9.20–15.20% (approx.
for the sample images) with the size¼100–20,000 pixel for
the analyses of the nuclei that has been adjusted. In this work-
flow, we have calculated the morphological information for
nuclei image data.
The integration of the information about the confocal
images is analyzed with the CellProfiler. We need to adjust
the image according to a preprocessing calibrating protocol.
This is a very important step, at the basis of quantitative mea-
surements. And then, by drag and drop the individual confocal
image we load the data into the cell profiler. The resulting
354 Garima Verma et al.