Status: Closed (View Workflow)
Affects Version/s: None
Fix Version/s: None
Sprint:Winter 1 Dec 28 - Jan 8, Winter 2 Jan 11 - Jan 22, Winter 3 Jan 25 - Feb 5, Winter 4 Feb 8 - Feb 19, Winter 5 Feb 22 - Mar 5, Winter 6 Mar 8 - Mar 19, Spring 1 2021 Mar 22 - Apr 2
Nearly every single-cell RNA-Seq pipeline aligns the "raw" sequence data onto a reference, usually a reference genome sequence. This is an essential step in the workflow. This step is done in order to generate "counts per gene per cell" spreadsheets that are then analyzed using unsupervised clustering methods and other such methods. The "counts per gene per cell" numbers are supposed to reflect the number of RNAs observed from the gene. However, there is a problem with that!
Because the methods for creating the sequence data start with minute amounts of RNA per cell, the protocols use PCR amplification in order to produce enough material for sequencing. This means that the same fragment of RNA will often get copied many, many times. Thanks to this, the number of sequences observed per gene will be only loosely related to the number of mRNAs from that gene that were in the original sample.
To get around this, the experimental protocols include a step that adds a "UMI" sequence tag to every read that came from the same RNA molecule. (UMI stands for "unique molecular identifier.") So instead of counting every single read, the data analysis protocols instead count the number of unique "UMIs" per gene. To keep track of UMIs in the data processing steps, the computational pipelines typically append the UMI sequence to the read name. In IGB, this "read name" is also the "id" attribute.
In addition to copying the UMI sequence, the pipelines also often copy another string that uniquely identifies the particular cell that the read came from. This is called a "cell barcode" and is also introduced into every read as part of the experimental protocol that produces the data.
Therefore in IGB it would be super-useful if we can create a filter that limits the reads being show to a specific string that the user enters. This would allow users to get a better sense of how often a given UMI appears in their data. It would also allow them to visualize gene expression for a single cell instead of looking at all of the data at once for every cell.
Also, this type of thing would be useful for any type of track, not only BAM tracks.
For this App, please implemented a new "filter by name" option that lets a user hide all items in a track that do not match the name.
Lastly, it is not clear yet whether we can introduce new filters as Apps. So a big part of this task will involve understanding the existing filtering system for tracks to see if an App can be added to implement a new one.