Synaptic Protein Zoo

We'll update this page as questions are asked!
If your question isn't here, check the Field Guide and Tutorials or ask us using the Talk boards!

General

Where can I see more examples?
You can find more examples by clicking on "NEED SOME HELP WITH THIS TASK?" and by visiting the Field Guide to the right of the screen.

What if I make a mistake?
No worries, we all make mistakes. Please read the tutorial carefully to avoid them and use the Field Guide and the "NEED SOME HELP WITH THIS TASK?" button to help you make your choices.

How do I know if I'm doing this right?
If some more examples would be helpful, you can find more examples by clicking on the Field Guide.

I made a mistake, can I go back and edit my classification?
Once you click on the "Done" or "Done & Talk" button, your classification has been submitted and stored in the Zooniverse's database where it cannot be edited. Before that point you can edit your submission. Trust your best judgement, and it's okay if you make a mistake every once in awhile. We have multiple volunteers review the same synaptic proteins and combine all these classifications together to make a final assessment.

Why is a human crowd needed to do this?
In general, an image contains about 500-1000 synaptic protein clusters. And we acquired about a hundred images and are still acquiring while you read this! Manually segmenting and classifying each cluster by a single expert is therefore too time-consuming!

Why can't computers do this?
While computer programs could significantly reduce the time needed to classify and segment synaptic protein clusters, they need clear instructions and a large number of examples to be able to learn how to do this. To obtain these insights and examples, we need to generate annotations.

I have a question or I found something interesting, who should I talk to?
The best way to reach the team is to post your question on the Talk discussion tool.

Segmentation

Both segmentations are not good: one too loose, one too tight. What should I do?
Select what you prefer. If the tight segmentation contains holes, select the looser segmentation. While you only see two segmentations, other options are available in other subjects. Therefore, it's normal that you sometimes have no choice but to pick a bad segmentation, and it's perfectly fine.

The image seems to contain two clusters, but the segmentation either combines both clusters or misses one of the two. What should I do?
Select the one that combines both clusters. It is more important to miss no cluster than to separate the two.

How to choose between two very similar segmentations?
Sometimes the two shown segmentations look alike and it can be difficult to choose between them. To help you out you can use use the + button to the right of the image to zoom in.

I still cannot choose between two very similar segmentations?
No worries, go with your gut feeling.

There is a hole in my preferred segmentation. Is that a problem?
Yes. If possible, choose segmentations without holes inside. Consult the "NEED SOME HELP WITH THIS TASK?" section and the Field Guide.

Both segmentations contain holes. What do I do?
Choose the segmentation with the smaller holes.

What should I do if I’m stuck between two segmentation images?
Please look at the two segmentations carefully, sometimes the difference may be very small (a view pixels). You can use the "Plus" sign right of the image to zoom in.

The subjects are displayed vertically instead of left to right. What do I do?
This issue happens either when the size of the window is too small on a desktop computer or when using the website instead of the app on a mobile device. On a desktop computer, try increasing the size of the window. On mobile device, ideally use the application. If this doesn't solve the issue, note that the images are displayed top to bottom instead of left to right.

Classification

What if I cannot clearly distinguish a cluster?
If you are not sure there is actually a cluster, please select "Other/Can't tell".

What if the cluster is outside the blue box?
Please only annotate clusters that are roughly at least 75% contained within the blue box. If the cluster is not within the box or only partialy within the box select "Other/Can't tell".

I am hesitating between two classes. What do I do?
Visit the Field Guide to the right of the screen to see more examples. There might be a cluster similar to yours. If not, go with your gut!

---

Didn't find an answer to your question? Visit the Talk section and ask the Synaptic Protein community!