Rukhsana Yeasmin Ph.D. in Computer Science Computer Science, Stony Brook University

RNAfolding Results for Hemoglobin Data

Experimental Results:

 Here, three graphs have been shown for minimizing the RNA Structure(or equivalently maximizing the bond energy) for Hemoglobin RNA of Frog. From the 1st graph, we see at first it maximizes the bond energy and goes upto 63. After that it starts increasing the bond structure. For a different starting of the structure, it continues maximizing the bond energy, finally reaches a stable state at 62.8. The 3rd graph shows the best result, it continues maximizing the bond energy, ending up at 4.3, this is also the final stable state for the structure. The final structure for this graph is as follows:

Now,

The first graph is showing the maximum bond energy for different iterations using different strategy. The black one is the graph for random change strategy. Here the algorithm just changes the structure randomly, even if it is stucked in a loop. The blue one is for the strategy, where we allowed neighbors of a particural codon to chnage even if the codon itself has been swapped with another codon from free codon table. Finally we used penalty function, whenever a repeated structure is found, for the next step we penalize each codon that had been changed in the previous step. The penalty is assigned for one step only. The magenta one indicates the performance of the algorithm with penalty function included.

 We see from the graph the Algorithm with penalty function performs better than other two. The second graph is for different penalty values. We used penalty values ranging from 1.0 to 0.3. Here we are multiplying the probability of change of the current codon (if this is a duplicate structure and the current codon was swapped in the previous step) with the penalty value. Hence, Lower the penalty value, more the codon is penalized. We see from the graph, penalty value .6 shows best performance. Penalty of 0.7 and 0.3 also show good performance.

Result 1 (Energy Vs Frequency Graph):

Number of Iterations vs Frequency Graph for each strategy:

Result 2 (Energy Vs Frequency Graph):

Number of Iterations vs Frequency Graph for each strategy: