Algorithm 1 Training GAN model
Input: Set of Sequences S, $ganCnt$
Output: GANs based sequences $S^{\prime }$
1: $m\_gen\leftarrow generator\left(\right)$	▹ generator model
2: $m\_dis\leftarrow discriminator\left(\right)$	▹ discriminator model
3: $m\_dis.compile(loss=CE,opt=ADAM)$
4: $seqLen\leftarrow len\left(S\right[0\left]\right)$	▹ len of each $S^{\prime }$ sequence
5: $iter\leftarrow 1000$
6: $batch\_size\leftarrow 32$
7: for i in $iter$ do
8:       $noise\leftarrow random(ganCnt,seqLen)$
9:       $S^{\prime }\leftarrow m\_gen.predict\left(noise\right)$	▹ get GAN sequences
10:     $m\_dis.backward(m\_dis.loss)$	▹ fine-tune $m\_dis$
11:     $m\_gen.backward(m\_gen.loss)$	▹ fine-tune $m\_gen$
12: end for
13: return($S^{\prime }$)