How to generate the Ground Truth?

[gaojunbin]

Dear authors,

Thanks for your excellent work. I have read your codes but not found how to generate ground truth file (e.g. ***.negative.bed.rdata, ***.positive.bed.rdata link, ***. grocaptss.bed link and so on). Can you give more information about the label generation process?

I also found that there are two similar functions (get_test_set and get_test_set0 link), It seems that get_test_set is loaded directly from the positive and negative label files (mentioned above), so I cannot get more details. Is the process of generating negative samples the same as get_test_set0 ?

Further, I found that most of the main programs used get_test_set but the parameters passed were based on the definition of get_test_set0. Is the code update not completely completed link?

Looking forward to your reply!

Best,
Junbin

[adamyhe]

Hi Junbin,

Apologies for the delayed response. I believe the negatives/positive datasets in that folder are defined following the methods section of the original dREG paper: https://www.nature.com/articles/nmeth.3329#Sec10

I'm uncertain as to where get_test_set0 is used, but it looks like the two function calls in the file you've linked use the correct get_test_set arguments (lines 65 and 87).