The question consult for author

[yaolp123]

Dear github authors:
recently I read your paper entitled"A novel end‑to‑end method to predict RNA secondary structure profile based on bidirectional LSTM and residual neural network",and it is a pretty good work. And I try to konw the Understanding data set, but have some doubt， described in part"Data collection and processing", for example, the zebrafish and yeast data sets , where the positive values indicated double-stranded bases and the negative values indicated single-stranded bases. And selected
the five lowest values as the single-stranded target bases, whereas the five largest values as the double-stranded target bases.
Then, 53964 RNA sequences why have 557425 target bases ? is not 53964*5=269820? And how about the human and mouse data sets , I have some understanding doubt in them. I am looking forward to your reply. Thank you for your nice work.

Best regards.

[qifei9]

Did you ever reached the authors?

Then, 53964 RNA sequences why have 557425 target bases ? is not 53964*5=269820?

I guess they selected 5 highest/lowest values and multiple bases can have the same value. For example, if the highest value is 7, then more than one bases could have that value/score.