Figure 1
The DeepSTF framework in general. (A) Data pre-processing: DNA sequence and shape profiles are transformed into feature matrix. (B) Sequence clade processing: capture higher-order characteristics in multidimensional sequence profiles. (C) Shape clade processing: extract possible characteristics from multidimensional shape profiles. (D) Output processing: feature mapping is transformed into the final prediction. (E) TF Module: the improved transformer encoder structure's comprehensive processing flow.

The DeepSTF framework in general. (A) Data pre-processing: DNA sequence and shape profiles are transformed into feature matrix. (B) Sequence clade processing: capture higher-order characteristics in multidimensional sequence profiles. (C) Shape clade processing: extract possible characteristics from multidimensional shape profiles. (D) Output processing: feature mapping is transformed into the final prediction. (E) TF Module: the improved transformer encoder structure's comprehensive processing flow.

Close
This Feature Is Available To Subscribers Only

Sign In or Create an Account

Close

This PDF is available to Subscribers Only

View Article Abstract & Purchase Options

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Close