Tissue Atlas: microscopy and annotation

Borbala Katona at the microscope

Today, we are back in the Tissue Atlas facilities at the Rudbeck lab in Uppsala. Borbala Katona and Maria Aronsson are research engineers in the group working with microscopy and annotation of stained tissues, which we described last week.

Borbala Katona has a bachelors degree in biomedicine and a masters degree in infectious medicine and has been working within the Human Protein Atlas since 2014.

Maria Aronsson has a masters degree in medical biology from Linköping University and joined the Human Protein Atlas in 2012.

Both Borbala and Maria are part of the Antibody approval, Protein Profiling, and Antibody Destiny-team, with Dr Cecilia Lindskog as group leader.

After immunostaining, the tissue microarray slides are examined under the microscope, and optimal antibody dilution and target specificity is assessed. The group uses available information in public databases on gene, RNA and protein level, internal and external RNA-seq data and internal technical validation in the decision process.

– For each antibody that we approve, a final immunostaining protocol is defined and applied to the full-scale tissue microarrays, Borbala Katona explains.

The images are then scanned for generation of high-resolution digital images.

Thereafter, the images are manually annotated, a job performed by both Borbala and Maria.

– We record the intensity and fraction of immunoreactive cells for each given cell population and determine the subcellular localization of the staining, Maria explains.

A text comment summarizing the characteristics for each antibody is added to the annotation. An independent second observer curates all finished annotations to ensure uniform annotations of high quality

After annotation and curation, protein profiles are evaluated and the generated data is compared with gene, RNA and protein characterization data. Antibodies that are approved at this stage are assigned a reliability score and a knowledge-based protein expression profile and are put in the queue for for publication in the next version of the Human Protein Atlas.

Maria and Borbala explain the extent of the job:

– If you do nothing else on one day, you can do 2-3 antibody annotations. There are very many images to look through, every antibody we stain with gives rise to 576 images that have to be evaluated manually!

Frida Henningson Johnson