Hi, I was wondering why ATF7 is the only TF out of the 32 listed in the data table that is not found in Factorbook. The same observation applies to the liver cell/tissue type, which is also not found in Factorbook. Is there anything special about this particular TF? Rick,

Created by Rick Farouni rfarouni
Every TF has some cell types as training data. See the table here https://www.synapse.org/#!Synapse:syn6131484/wiki/402034 . You can use the training cell types for each TF to train models (e.g learn motifs etc.) for that TF. Then you need to predict in the ladderboard test cell types and in the final submission test cell types. -Anshul.
Hi Anshul, Thanks for the reply. This is what I thought. So in total, the incomplete data is for 32 TFs and 14 cell types/tissue, but only 31 TFs and 11 cell types/tissue can be used for training. Rick,
There are ATF7 motifs available in CISBP as well http://cisbp.ccbr.utoronto.ca
See the Resources section https://www.synapse.org/#!Synapse:syn6131484/wiki/402036 . I lit several popular motif discovery tools. Eg. MEME-ChIP which is part of the MEME suite of tools
Can you please recommend a motif discovery tool? I am not familiar with these tools. Thank you
Why do you need a previously identified known motif? We provide training ChIPseq data for every TF. You can easily learn a motif from it using your favorite motif discovery tool.
Hi Majia, thanks for noting that, this is an important point. Please check out the [challenge resources](https://www.synapse.org/#!Synapse:syn6131484/wiki/402038) for additional motif databases, you will find a PWM for ATF7 for instance in jaspar. Cheers, Robert
But if it is not existing in Facebook, then how can we acquire the Position Weight Matrix to scan the motif through the genome?
Hi Rick, ATF7 is a newly generated sample. It is not present in FactorBook because no ChIP-seq datasets were previously available for it so its never been analyzed before. Same applies for all the true-blind hidden datasets e.g. liver samples. This is by design. These are hidden test sets so we can test true generalization performance of methods on unseen datasets. Thanks, Anshul.

ATF7 in data table but not in Factorbook page is loading…