satoken
satoken
Hi, Thank you for using MXfold2. Please convert 'T' to 'U' and try again. The web server does this conversion automatically. Kengo
MXfold2 works well for short sequences (30~50bp). However, I don't think it is very accurate for very long sequences (5000 ~ 11000bp), although it can be tried if memory and...
It is the folding score, which corresponds to Eq. (1) in the paper (http://dx.doi.org/10.1038/s41467-021-21194-4). A large value means a stable structure.
The default model of MXfold2 was trained to be as close as possible to the sign-reversed value of the thermodynamic energy using thermodynamic regularization. See Table 3 in the paper...
You can input as long an sequence as your memory allows.
Unfortunately we have not implemented batch training because it is difficult to align the length of the sequences in a batch.