Session Transcript for User A

Friday, March 3, 2006

Experience with IMG: Novice.
Experience with genes: Extensive.
Experience with annotations: Extensive.
User A has previously worked at GenBank, so he feels quite confident doing batch annotations.

TASK 1: Look at a specific gene's annotation history

With some hesitation, User A attempted to locate the 'annotation history' through 'update annotation' (going to the Modify Annotation page) before finding 'annotation history.' He also attempted to click the annotator's name, but this only popped up an email client.

TASK 2: Modify an existing annotation with homologous data

User A returned to the Modify Annotation page, but was somewhat confused about where to get all the information about a homologous gene of interest. He searched for a phylogenetic tree and for BLAST hits, but could not find the BLAST hits. "I think I might be missing it?" He makes his way back to the Gene Details page, where he clicks on NCBI Blast - he does not notice the homologous genes, because they are at the bottom of the 'page' which is falling off the table.

He starts entering the data in Modify Annotation, but when he gets to the pre-fill drop-down menu, he selects the second homologue which autofills the data. “I guess I believe it,” he says, but to verify he explores the link to the GO Function, Malate Dehydrogenase Activity and the link to the COG group, Malate/Lactate Dehydrogenase.

He further explores the COG data, identifying a COG with 377 genomes. He clicks on the 377 link, which takes him (somehow) to the gene search page. The results look promising, so he is satisfied with his modification.

TASK 3 : Save a draft of this annotation modification

He clicks 'save draft' and goes to the 'preview' page. He looks for an OK or 'Accept Annotation' button, but cannot find it. He decides to click on MyAnnotations. From the list of drafts, he clicks 'submit.'

TASK 4 : Express opposition to someone else's choice of annotation

He clicks the 'I Disagree' button, and jokingly suggests a giant slash through the existing gene annotation.

TASK 5: Change an annotation by adding a COG group to it

From the Gene Cart, he clicks Update. The Modify Annotation window pops up. He clicks COG Group (which opens up the COG Browser). He then goes to the Prefill Values drop-down box and selects a different homologous gene. This changes a bunch of values in the annotation, not just the COG group! He saves a draft, goes to MyAnnotations, and submits. (I.e. he does not notice the ‘Submit’ button on the Modify Annotations page, since he is already familiar with the Save Draft button.)

Possible solutions

User A says: "This is cool." The Phylogenetic Occurence Profile link from the Gene Details page is a good way to study what genes are in which organism, and what needs to be cut.

The most useful tools for annotating block genes are phylogenetic trees, multiple alignments, and phylogenetic occurrence profiles.

Historical Interlude from User A

N at GenBank was adamant about taking this annotation power away from people This is why there are so many checks in the current GenBank system, to confirm vector screens and cut up genes in interesting ways. The other problem is gene families; genes are called part of a gene family, but sometimes they're paralogs, not orthologs.

 

Open Questions

What do you think of history/batting average?

User A likes knowing the names of the last annotators.

Batting average would be good, although he wonders what the repercussions of competitiveness are. He believes the annotator's name has more credibility than the batting average.

Would you actually click "I agree"?

If I agreed, I would probably leave the annotation untouched. (No) A disagree is an active thought. An agree is passive.  

What is a defLine?

DefLine uses a very specific syntax to draw together data from different fields and check them against each other. It consisted of the name of the organisms, the gene name, gene/MRNA, and complete/partial.

Would a multiple annotation tool be useful?

Definitely. Annotators could get 300 KB of unannotated sequence at a time, or a massive phylogenetic set, where they did 3000 HIV isolates from 20 people. Moreover, for a comparative tool, you would have more people doing multiple annotation .

What do you think of the 'fellow' system?

"That's a Good Idea"

There will be lots of gene-specific fanatics. You could specifically invite reputable individuals to become automatic fellows in this system.