Skip to main content
Premium Trial:

Request an Annual Quote

Caltech Researcher Says Digital Organisms Could Aid Bioinformatics Development


The phrase “in silico biology” is thrown around quite a bit these days, but Chris Adami means it quite literally.

Adami heads up the Digital Life Laboratory at the California Institute of Technology. The team has developed software called Avida — an environment in which populations of self-replicating computer programs can live, evolve, and adapt.

The software was developed with the intent of studying the mechanisms of evolution; thus, it differs from traditional genetic algorithms whose “evolution” is based solely on externally imposed fitness functions, or user-defined “traits.” In Avida, the digital organisms must interact with each other in order to reproduce.

A typical Avida organism is a single “genome,” composed of a sequence of instructions that represents several hundred equivalents of amino acids. A generation spans a second or less, compared to around 15 minutes for E. coli. Thousands of generations can be run in days or weeks, rather than years, and every step of the process is saved on disk.

Adami has a background in theoretical physics, and developed the Avida software to find the “underlying principles” of Darwinian evolution. But it’s not all blue-sky research, according to Adami. There’s an immediate — and very practical — application in bioinformatics: “It’s very useful for testing bioinformatics algorithms,” he said.

Many bioinformaticists use “synthetic data” to evaluate the accuracy of their algorithms, but Adami said that data derived with the Avida software would be more effective at assessing the strengths and weaknesses of new programs, particularly in the areas of phylogenetic tree reconstruction and microarray analysis.

“In digital life, we have strong control over both sequence and function. While we evolve sequences, we’re able to determine what their function is,” Adami said. “The advantage is that you could treat the sequence data that we generate as you would treat any sequence data, and apply the bioinformatics algorithms on them, and then compare it to what we actually know their function is.”

This approach would help bioinformaticists determine whether there are certain circumstances under which aspects of function are missed, or whether there are regions of code that are more vulnerable to mutations than others and how that would affect an algorithm, he said.

Charles Ofria, an assistant professor at Michigan State University and a former grad student under Adami, is currently using the approach to evaluate algorithms for phylogenetic tree reconstruction.

Current methods of evaluating these programs “assess about how well the tree represents the data, but don’t assess how well the tree you get represents the actual tree — the actual relationships between the sequences — because you don’t have the historical data.”

Using Avida, “when we evolve our populations, at the same time we can keep track of the actual [phylogenetic] tree, and therefore if we have a number of sequences that we want to get the ancestral relationships between, we could just use one of these algorithms to reconstruct it and then compare this reconstructive algorithm to the actual tree,” Adami said.

Adami was quick to note that the digital organism approach isn’t applicable to all types of data that bioinformaticists are interested in. For example, he said, because the mutation rate per gene is much smaller in higher organisms, any effect that relies on a very small mutation rate per gene would be difficult to see using digital organisms. Another drawback: “We don’t have transcription and we don’t have a translation stage. So anything that is directly related to splicing is not going to be captured by the digital approach.”

Finally, he said, “until we implement some possibilities of gene regulation and co-regulation into the system, I don’t think we’ll be much help to the people investigating gene networks.”

However, Adami noted, Avida does produce one thing that every bioinformaticist will appreciate: “We create a wealth of data. And in order to get sequence data, we can use the print command, we don’t have to use PCR.”

The Avida software is freely available at:

— BT

Filed under

The Scan

Shape of Them All

According to BBC News, researchers have developed a protein structure database that includes much of the human proteome.

For Flu and More

The Wall Street Journal reports that several vaccine developers are working on mRNA-based vaccines for influenza.

To Boost Women

China's Ministry of Science and Technology aims to boost the number of female researchers through a new policy, reports the South China Morning Post.

Science Papers Describe Approach to Predict Chemotherapeutic Response, Role of Transcriptional Noise

In Science this week: neural network to predict chemotherapeutic response in cancer patients, and more.