Tweet
An accumulation of somatic mutations in genes necessary for cell growth, regulation, and division can often lead to cancer and reinforce cancer phenotypes. Protein-affecting mutations (PAMs), gene fusions, and copy number alterations (CNAs) are the three primary somatic mutations that can alter gene functions or cause a non-functional gene. The mutations by PAMs, gene fusions, or CNAs typically have similar effects on cancer phenotypes, referred to as allelic heterogeneity.
Most studies linking somatic mutations to cancer phenotypes only focus on one type of mutation and consequently miss certain links between other mutations. Additionally, focusing on a single mutation type reduces the ability to identify associations for mutations with high allelic heterogeneity.
In a new study published in Cell Reports Methods, a team of researchers combined the analysis of all three mutation types to better capture the impact of somatic mutations on cancer phenotypes. The work led to the development of a new software tool called OncoMerge. This tool is able to detect PAMs, gene fusions, and CNAs and then analyze the system behind the mutations, revealing the connections that can be used to build a model for predicting future changes caused by these mutations.
Discovering patterns in data
“We are able to look at the gene expression patterns using correlation,” said Christopher Plaisier, the team’s leader and an assistant professor for the Ira A. Fulton Schools of Engineering at Arizona State University. “Then we can see what is being activated or repressed, which allows us to look at the deeper functions behind that.”
The idea of OncoMerge has been with Plaisier since his postdoctoral work when he realized the need for a tool that could be used to analyze the network behind these mutations. Plaisier then leveraged his experience in human genetics, cancer biology, and computational biology to begin developing his idea.
During the design process, Plaisier’s team verified that certain feedback networks were boosted to create an abnormal regulation of networks, demonstrating that the tumor can control its environment in order to sustain itself. After applying OncoMerge to over 9,000 patient tumors, the team validated their methods and confirmed that integrating mutation data improved the accuracy of predictions for linked behavior among genes.
Plaisier and his group hope that OncoMerge will be integrated into other pipelines to improve downstream analyses linking somatic mutations to cancer phenotypes.
Improving Data Science Education
In addition to developing tools for cancer modeling, Plaisier is interested in improving data science education by calling for more introductory coding courses to better familiarize students with coding techniques. Plaisier also wants to create a data science elective course for biomedical engineering students.
“Seeing how cancer biology and bioinformatics interact has taught me the importance of applying lab skills and computational skills to ask and answer interesting questions,” said Sierra Wilferd, a graduate student in Plaisier’s lab, who appreciates her advisor’s interdisciplinary approach.
Understanding complex systems like tumors and other diseases require a multifaceted approach like the one utilized by OncoMerge. As researchers continue to develop and apply these types of software tools, they will be able to make the needed discoveries for improved clinical outcomes.
Most studies linking somatic mutations to cancer phenotypes only focus on one type of mutation and consequently miss certain links between other mutations. Additionally, focusing on a single mutation type reduces the ability to identify associations for mutations with high allelic heterogeneity.
In a new study published in Cell Reports Methods, a team of researchers combined the analysis of all three mutation types to better capture the impact of somatic mutations on cancer phenotypes. The work led to the development of a new software tool called OncoMerge. This tool is able to detect PAMs, gene fusions, and CNAs and then analyze the system behind the mutations, revealing the connections that can be used to build a model for predicting future changes caused by these mutations.
Discovering patterns in data
“We are able to look at the gene expression patterns using correlation,” said Christopher Plaisier, the team’s leader and an assistant professor for the Ira A. Fulton Schools of Engineering at Arizona State University. “Then we can see what is being activated or repressed, which allows us to look at the deeper functions behind that.”
The idea of OncoMerge has been with Plaisier since his postdoctoral work when he realized the need for a tool that could be used to analyze the network behind these mutations. Plaisier then leveraged his experience in human genetics, cancer biology, and computational biology to begin developing his idea.
During the design process, Plaisier’s team verified that certain feedback networks were boosted to create an abnormal regulation of networks, demonstrating that the tumor can control its environment in order to sustain itself. After applying OncoMerge to over 9,000 patient tumors, the team validated their methods and confirmed that integrating mutation data improved the accuracy of predictions for linked behavior among genes.
Plaisier and his group hope that OncoMerge will be integrated into other pipelines to improve downstream analyses linking somatic mutations to cancer phenotypes.
Improving Data Science Education
In addition to developing tools for cancer modeling, Plaisier is interested in improving data science education by calling for more introductory coding courses to better familiarize students with coding techniques. Plaisier also wants to create a data science elective course for biomedical engineering students.
“Seeing how cancer biology and bioinformatics interact has taught me the importance of applying lab skills and computational skills to ask and answer interesting questions,” said Sierra Wilferd, a graduate student in Plaisier’s lab, who appreciates her advisor’s interdisciplinary approach.
Understanding complex systems like tumors and other diseases require a multifaceted approach like the one utilized by OncoMerge. As researchers continue to develop and apply these types of software tools, they will be able to make the needed discoveries for improved clinical outcomes.