Data Analysis and Annotation

Maintaining a high-level of quality and accuracy in our data and providing valuable data analyses requires deep expertise in neuroanatomy, gene expression data, and genomic analysis.

The data analysis and annotation team includes experienced Ph.D.-level neuroanatomists and a diverse group of accomplished data analysts who are highly skilled in interpreting gene expression data.  With expertise from mouse to human neuroanatomy, we can analyze accurately any gene expression characteristic in more than 1,000 brain structures including the olfactory bulb and spinal cord, and identify specific cell types, such as different classes of neurons, glia and endothelial cells.  In addition, data across different individuals or groups can be analyzed to assess changes in gene expression according to species, disease states, ages, or gender.

Data analysis is a critical step in the workflow for the creation of all of our online public resources.  To ensure the accuracy of data collection, our analysts carefully review each tissue section after microscopic imaging and before images are released publicly. Strict, standardized quality control guidelines, which differ slightly for each project, provide a framework for analysis.  Not only is this quality control step necessary for accepting data for inclusion in our online resources, but it also provides essential feedback to the data production teams in our high-throughput laboratory about recurring artifacts or changes in data quality that warrant investigation, as well as improvements to the data production processes.

In addition, data is often further characterized, or annotated, beyond initial quality control processes in order to improve the accessibility and value of our data to end users.  Images are examined in detail to assess, for example, where and with what degree of regional specificity a gene is expressed, and the intensity level and local spatial pattern of its expression.  Such analysis formed the foundation of the Expression Categories search in the ALLEN Spinal Cord Atlas.  Our manual data analysis efforts are often coupled with informatics analysis, to create robust data mining tools—such as the NeuroBlast and Fine Structure Search functions in the ALLEN Mouse Brain Atlas—that combine the power of automated algorithmic searches with the benefits of manual curation.

Lastly, data analyses also support efforts beyond our public resources themselves.  Some analyses are done in collaboration with investigators at other institutions to support their studies; others are used for piloting a new approach to data collection; and others are used internally by our neuroscientists and support publications in peer-reviewed scientific journals.