CalceinHisto

‌CalceinHisto measures bone formation parameters from samples that have undergone in vivo double labelling with Calcein. The bone is stained fluorescent blue using a Calcein Blue stain (see staining protocols for details). The slides are then imaged on a fluorescent microscope in the blue and green channels and the images saved as dual band tiff images. Standard colour images can also be used as the input but the software may not identify the two colours the correct way round. For reliable analysis of double label distance a pixelsize of around 0.5µm or less is required for mouse bone.

Setup

The routine info dialog box allows setting of the size calibration as well as the label interval in days. As there are two fluorescence channels, it can toggle between the display of what is shown below

 

Identifying Bone:

 

Identification of bone is by a double (or hysteresis) threshold. The inner threshold identifies parts that are very clearly bone. These regions are then grown to the second (less stringent) outer threshold. This allows the “filling out” of the bone area to its lighter stained edges without background coming up.

 

Finding the calcein Labels:

In the next step, the software finds the centre line of bright, line-like structures. This is not just a threshold on brightness!! The main controls to adjust are the inner and outer threshold settings.  Increasing the “minimum branch-length” can prune short branches, and “smooth centre line” will do what it says. However, high values of the smoothing may lead to double labels turning into single labels! The pre-filter radius and Gaussian sigma should not normally be changed. The results can be edited manually using the drawing controls.  

 

Classifying the labels:

In the final step the labels are classified as either double or single labels. Main parameter here is the “maximum double label distance”. This determines how far from each label the software will look to see if there is another label to form a double label. This step normally requires some manual editing as the current algorithm is not capable of distinguishing double labels and single labels on both sides of thin trabeculae. Hopefully this will be improved in the next version.

Results: