Penn State researchers have developed a new computer program that helps to simplify steps in predicting the outcome of DNA shuffling.

The process has the potential to help industries in improving enzymes, vaccines, and therapeutic proteins.

The shuffling process fragments two or more genes into smaller pieces, splits the strands and then recombines the fragments into new DNA sequences through a procedure called annealing, which joins the single-stranded DNA through cooling.

The original DNA double strands are randomly cut into pieces through an enzymatic process called nuclease treatment, said Stefan Lutz, postdoctoral fellow in chemistry. Lutz worked with the experimental data during the research.

When the double helix is split into single strands, the DNA is separated from its complementary parental sequences and searches to anneal with a similar yet distinct new combination.

A section of the single strand from one of the original genes can anneal with a strand from the other original gene; the portions of each strand that did not anneal find their complementary sequence through the process of polymerase treatment.

The point at which the annealed sections end and the unannealed sections begin is known as the crossover point. The crossover points allow for strands of different genes to combine. This process is then repeated many times to create potentially millions of combinations between the original genes, Lutz said.

The goal is to produce a new DNA sequence that codes for a better protein than either of the original sequences would.

"Primarily what we're trying to figure out is evolution in a test tube to create better proteins," said Costas Maranas, assistant professor of chemical engineering.

The goal of the computer program is to model the chopping up and reassembling process into a simplified time-efficient manner.

When the possible combinations reach into the trillions, it is difficult to research and test all of the modified genes for improved proteins, Lutz said. The computer program condenses the numerous possibilities into a smaller number that is easier to examine.

"It allows you to play a number of scenarios on the computer so that you don't have to waste experimental resources," Maranas said.

The computer algorithm inputs the DNA sequence of the parental genes as well as thermodynamic data and predicts the probability and position of crossovers.

The program has been tested, and it produces comparable results to the previous model.

"The program makes the procedure take two cycles instead of 10. It saves time, effort and expensive equipment," said Greg Moore (graduate-chemical engineering) who has been involved with the program for one year.

The research was published last week in the Proceedings of the National Academy of Sciences.