The i5k Workspace/Web Apollo 'Replaced Models' field - FAQ

How the i5k pilot/i5k Workspace is handling OGS merges - the new ‘Replaced Models’ field in Web Apollo

 

  • Background: What is an OGS?

    • An OGS is an ‘Offical Gene Set’ - a single set of gene models that is the official representation of the gene content of your genome assembly.  

 

  • What is an OGS merge?

    • This is the process of taking your genome assembly’s manual annotations from Web Apollo and combining them with your computational gene predictions (e.g. from MAKER) to generate a single, non-redundant OGS.

 

  • Why is an OGS merge important or useful?

    • It’s important if you want to have a single, non-redundant gene set to distribute to your research group and the scientific community at large that reflects the efforts and improvements of your manual annotators.

 

  • How does this new ‘Replaced Models’ method work?

    • It’s pretty simple. First, make sure you know what gene set you’re annotating (usually, there will be one set of computationally predicted genes listed under the track category ‘Primary gene sets: protein coding’, e.g. LDEC-v0.5.3 - ask us or your genome coordinator if you’re not sure). When you’re finished with an annotation, in the Information Editor, go to the ‘Replaced Models’ section. Click ‘Add’, and under ‘Transcript name’, type in the name (or ID) of the computational model your annotation replaces. See our tutorial for more details and examples.  

    • The information that you enter in the ‘Replaced Models’ section tells us which computationally predicted models should not be included in the OGS - usually because they will be replaced by your manually curated model. At the end of your group’s annotation period, we will search for the ‘Replaced Model’ information. All of the models listed under ‘Replaced Models’ will be removed from the OGS, to be replaced by your manually curated models. See the tutorial for examples.

 

  • Does this new ‘Replaced Models’ rule apply to my group?

    • If you want us to handle the generation of your official gene set (OGS), yes. If you have somebody else lined up to do this, no. Please let us know if you already have someone doing this.

    • We can only generate your OGS if you’re following a fairly simple model - you are manually curating only one computationally predicted gene set (e.g. LDEC-v0.5.3).  There are much more complex ways to generate an OGS, and you’re welcome to go down that path - but we won’t be able to do this for you.

 

  • Why are you implementing this? Can’t you figure out what models to replace computationally?

    • Sure - maybe 95% of the time. The remaining 5% is pretty hard to get right. In those cases, though, it’s quite straightforward for the annotator to see which gene models from the computational predictions don’t belong in the OGS. (We have a tutorial with some examples to show you how).

 

  • What do I do if my group is really far along in the annotation process? My annotators are going to revolt!

    • If your group has already started the annotation process, we will provide you with a spreadsheet that lists all of the models your group has created, including a URL. That way, each annotator can review their models and add the requested ‘Replaced Models’ information.

    • If one of your annotators has too many models to go back to, they don’t have to go back to the Information editor - they can return the spreadsheet to us with the ‘replace’ information in the final column.

 

  • What happened to all of the ‘canned comments’? I was using them to include this information already!

    • Yes, and we thank you for it - but annotators weren’t doing so consistently, which makes it difficult for us to process computationally. The new ‘Replaced Models’ section is now a standardized method that we’re enforcing across all manually curated models.

    • You’re still welcome to use the ‘comments’ section to share other, custom information about your manual curation with the research community.

 

  • I still don’t understand this.

    • See our tutorial for more details and examples.

    • Contact us if you have any more questions.