In Huffman Coding we create a new node by combining the two symbols with least frequency(or probability in this problem) present in the data. The value of the new node will be equal to the sum of the probabilities of the nodes combined. Now, this new node created will be viewed as a new symbol in the data with the calculated probability. While calculating the next New Node we consider this node also.If it is one of the two symbols present in the data with the least probabilities then it is used to calculate the next New Node else the other corresponding symbols are used. Hope this helps :)