We are thankful for the hard work you put into the ALS Stratification Challenge and excited by the results we have seen. We look forward to others building on your work, using your Synapse-based write-up and source code as reference material, linked from the Challenge results page:
Along with encouraging ongoing ALS research, we must also do our diligence to control access to the underlying, sensitive human data, per the PRO-ACT Terms and Conditions (T&C). In particular, the T&C that you signed with PRO-ACT states that the ALS2 Challenge Data sets -- PRO-ACT as well as the two registry datasets – can only be shared with individuals that have signed the T&C.
To ensure that all of the participant projects listed on the public leaderboard are in compliance with the T&C, we are taking the necessary step of temporarily blocking file download for users other than the project owner, and asking for each Challenge participant who submitted a final submission to confirm back to us in writing at firstname.lastname@example.org that the contents of their project have been checked and contain no Challenge data. Included in your email confirmation, please provide us with your Synapse User name and a link to the shared Synapse project in question. Once we have received your written confirmation, file download will be unblocked.
In the event that your source code requires the Challenge Data we asked you to remove, please add a statement to your Project Wiki on how an interested user could -- after registering and obtaining the Challenge Data from PRO-ACT -- restore an executable version (perhaps also include a processing script in case your submission requires the training data in a specifically processed version).
We want to make sure you know that your compliance with this step is required in order to be listed as an author on the ALS2 Challenge paper.
Once again, thank you for your support.
The ALS Challenge Admin Team
Help get this topic noticed by sharing it on Twitter, Facebook, or email.
I would like to clarify exactly what do we need to remove. Is it any of the raw data files, or any data file that has been derived from the original data as well?
One example: In order to cluster patients I have a file containing the training data represented by 6 variables, scaled, and without any header indicating what the variables are, neither patient IDs. It is just a bunch of numbers:
Does this still break the T&C?
What about a file having the mean & deviation of some of the variables for the training set?
can you give me the leaderboard and validation data?Because I am concerned about the challenge of the time is relatively late, so did not submit the final program. but now,Compared to the final results (team:UglyDuckling, ZhuLab..), I found that my experimental results are also relatively good,I hope you can give me validation set, I would like to verify my approach.