In silico nano-dissection webserver

Cell-lineage-specific transcripts are essential for differentiated tissue function in metazoan organisms. They are frequently found to be the cause of hereditary disease and mediate progression of acquired diseases. Identifying the tissue specific transcriptome can guide disease gene identification in genetic studies and the development of organ specific therapeutic targets. This server performs an in silico nano-dissection, which is an approach we developed to identify genes with novel cell-lineage specific expression. This bioinformatics strategy leverages high-throughput functional genomics data from tissue homogenates to accurately predict genes enriched in specific cell types.

Nano-dissection can define the cell-type-specific transcriptome from comprehensive tissue gene-expression datasets. This webserver provides datasets for commonly studied tissues, and we are happy to add additional datasets upon request.

Creating a nano-dissection

The first step is to click the "Start Nano-dissection" button which will allow the user to choose a title and description for this nano-dissection. To make the nano-dissection public (viewable to all visitors) check the "Public" box.

Choose a Dataset

The next step is to choose the dataset to use for nano-dissection. We recommend choosing the most specific dataset that includes your cell lineage of interest. Available datasets include hand curated subsets of GEO, a complete GEO collection and renal-microdissection dataset used in our manuscript.

Add Gene Sets



Next, sets of genes must be added to the nano-dissection. These gene sets can either be public (from HPRD or created by other users) or the user can create a custom set. Gene sets are either positive (consiting of genes known to be specifically expressed in the cell lineage of interest) or negative. The best negative genes are those known to be expressed in nearby or physically touching lineages, but if these are unavailable, random negatives may also be used. For this algorithm postive sets define the lineages of interest while the negative sets define the desired specificity.

Custom Gene Sets

In many cases, no public gene set of the desired cell lineage(s) will be available. A user can define his or her own gene set using unpublished results or a thorough literature search. Immunohistochemical stainings from the Human Protein Atlas may also be helpful to define custom gene sets for under-annotated lineages. Once the appropriate genes have been identified, they are pasted or typed into the text-box. This webserver will attempt map genes. The user must then select the appropriate gene if there are ambiguous matches and save the set. Once

Advanced Iterative Approach

Often biologists will know that some standards (either custom created, or from a database) are of higher specificity than other standards. For example, a standard made up of cell-lineage specific genes identified through double immunoflourescence would be of very high specificity, while one pulled from the public database may be more variable. Nano-dissection is capable of identifying and using only the standards with most appropriate level of specificity (for the relevant decision rule, see the nano-dissection manuscript).

Enabling Advanced Features

The advanced view allows a user to specify the appropriate level of specificty for each standard.

Labeling Specificty Levels

The drop-down boxes allow each standard to be assigned a relative specificity level. Nano-dissection treats these as relative (i.e. the ordering should be right, but no absolute specificity must be known). This information allows nano-dissection to best use your domain specific knowledge to identify cell-lineage specific transcripts.

Results View (Part 1)

As with the simple view, the results view shows the predicted value of all genes in standards and top novel candidates are presented. Interactive figures allow the user to turn on and off display of each standard by clicking on the standard name in the legend.

Results View (Part 2)

At the bottom, the standards are listed. Standards shown with a green background were used for the analysis. Standards with a white background were identified as unnecessary or harmful by the nano-dissection decision rule and were not used for the final predictions shown above.