You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello,
I have a model that predicts TF binding for a mixture of several possible TFs (the data is similar to having peaks from ATAC-seq). I have computed actual and hypothetical importance scores for only the True positive predictions. I wish to know which settings would be recommended that I change for modisco.
-sliding_window_size: I currently leave at 15 because I can't know exactly what the core motif sizes are. (BPNet paper used 21 but I think the core motif sizes probably had some known estimated sizes)
-target_seqlet_fdr: I am not sure how to change this. Any suggestions? (BPnet uses 0.01 but the example notebook uses 0.15, so how does one make a decision on which value to use?)
And are there any other settings are recommended I change since I have several possible binding motifs?
cheers
Fritz
The text was updated successfully, but these errors were encountered:
Hello,
I have a model that predicts TF binding for a mixture of several possible TFs (the data is similar to having peaks from ATAC-seq). I have computed actual and hypothetical importance scores for only the True positive predictions. I wish to know which settings would be recommended that I change for modisco.
-sliding_window_size: I currently leave at 15 because I can't know exactly what the core motif sizes are. (BPNet paper used 21 but I think the core motif sizes probably had some known estimated sizes)
-target_seqlet_fdr: I am not sure how to change this. Any suggestions? (BPnet uses 0.01 but the example notebook uses 0.15, so how does one make a decision on which value to use?)
And are there any other settings are recommended I change since I have several possible binding motifs?
cheers
Fritz
The text was updated successfully, but these errors were encountered: