A nano/micropatterned nuclear deformation based cellular diagnostic system

A diagnostic method for a detection of a nuclear deformation based a cellular differentiation, comprising the following steps:

• a) preparing nano/micropatterned surfaces:
• 1) placing a silicon wafer produced by a photolithography or an ion beam lithography into a container,
• 2) pouring poly(dimethyl siloxane) (PDMS) onto the silicon wafer and curing the silicon wafer with the PDMS in an oven,
• 3) solidifying a PDMS mold during curing, and peeling off the PDMS mold from the silicon wafer,
• 4) using a PDMS replica as a frame, and
• i) pouring a biodegradable polymer solution or a nondegradable polymer solution into the PDMS replica, and evaporating a solvent, or
• ii) hot embossing the PDMS replica with a pre-prepared polymer film, or
• iii) directly on the silicon wafer,
• 5) performing a crosslinking step on polymers requiring a chemical or photocrosslinking;
• b) performing a cell seeding on the nano/micropatterned surfaces;
• c) imaging:
• putting the nano/micropatterned surfaces with cells on a top as samples for an imaging,
• staining each sample of the samples with a nucleus or a DNA dye, and
• imaging the each sample with a confocal laser scanning or a fluorescence microscope;
• d) analyzing nuclear deformations of a single cell and a cell population with an algorithm method comprising the following steps:
• reading an image in a RGB format,
• converting the RGB format to a gray scale,
• detecting cell nuclei in the image,
• excluding cell nuclei with low signal intensity,
• getting binary images using an Otsu method,
• excluding cell nuclei at four borders of the binary images,
• excluding very large and very small nuclei,
• detecting cell nuclei going through a cell division with a watershed transformation, and excluding cell in mitosis,
• after a quality control, smoothing remaining cell nuclei images by using a Gaussian filter to even rough pixel edges,
• applying morphological to eliminate any artefacts,
• rotating each nucleus by orienting a major axis of the each nucleus to a y-axis, and ensuring a centroid is always on a right side of the major axis,
• for a scale invariance, scaling a longest dimension of an image of the each nuclei to 64 pixels while preserving an aspect ratio,
• cleaning a noise with a low cut filter, and/or a morphological opening, and/or a median filter,
• excluding nuclei that occupy 100% of bounding boxes of the nuclei, and the nuclei are clearly artefacts,
• calculating a rectangularity (R) of the nucleus:
• R
• =
• 
• π
• 4
• -
• A
• cell
• A
• bounding
• 
• 
• box
• 
• calculating a circle variance (CV):
• μ
• R
• =
• 1
• N
• 
• ∑
• i
• =
• 1
• N
• -
• i
• 
• d
• i
• ,
• 
• σ
• R
• =
• 1
• N
• 
• ∑
• i
• =
• 1
• N
• -
• i
• 
• (
• d
• i
• -
• μ
• R
• )
• 2
• ,
• 
• CV
• =
• σ
• R
• μ
• R
• calculating a deformation score (DS).