The identification of cell-type specific or brain-region specific transcripts adds specificity for the diagnosis of a wide variety of brain disorders, such as traumatic brain injury or brain cancers. One goal of identifying cell-type and brain-region specific genes is that they encode for proteins that could be found in the blood – providing potentially highly specific information about the spatial location of perturbed networks from a non-invasive proxy measurement. The development of such multi-parameter blood protein markers lies at the very heart of the predictive medicine. Thus, in this study, we have proposed a novel method - called EigenBrain - to identify novel candidate cell-type specific transcripts, and validated its performance using 10-fold cross-validation. All cases studied except the coronal section of neuron cells showed more than 95% specificity and 95% accuracy. Our EigenBrain approach transformed the original feature vector into the new lower-dimensional EigenBrain space, which greatly improved the accuracy of our algorithm for identification of cell-type specific transcripts. We applied this approach to the set of in situ hybridization (ISH) mouse brain images of 20,000 genes in the Allen Brain Atlas dataset and discovered a strong candidate set of brain-region and cell-type specific transcripts. Furthermore, we also investigated these highly expressed patterns in brain regions for each cell type - called EigenBrain. The enrichment of functional categories and relevant pathways involved in each candidate cell-type specific transcripts were investigated for further biological validation.
You will be able to download and print a certificate for these PDH credits once the content has been viewed.
If you have already viewed this content,
please click here