ILLINOIS — A desire to improve the status of cancer patient treatments led former theoretical physicist, Jun Song (PhD), to pivot his career towards cancer research. He spent several years learning about cancer biology and retraining to apply his background in high-energy physics to cancer-related high-throughput genomics analysis.

This type of analysis uses sequencing assays to infer the geometry of DNA as it wraps around proteins and folds into a condensed structure called chromatin. Depending on the arrangement of the DNA, certain areas of the genome are made accessible or inaccessible to other proteins in the cell.

“I was interested in the geometry of genomic information and how that controls the expression level of a gene,” said Song, Founder Professor of Physics and Cancer Center at Illinois member.

This is especially important for cancer research as some of these genes may be oncogenes, which can cause a cell to develop into a cancer cell, and tumor suppressor genes. Many of these genes can be activated through epigenetic mechanisms, which encode information beyond the DNA sequence content.

As a physicist, Song contributes to cancer research through signal processing. Genomic coordinates embody multi-dimensional data with information about DNA sequences, epigenetic modifications, protein binding, and more. However, these measurements can be noisy, containing biases or incorrect information; Song’s role is to remove this noise and extract the relevant information for cancer researchers.

“You can summarize what I do as signal processing using probability theory and statistics, using tools from physics and mathematics to model the distribution of proteins on DNA and study the structure of biological data,” said Song. “My lab is also developing tools to extract biologically relevant features implicitly learned by machine learning algorithms and AI.”

These techniques may be applied to the diagnosis and treatment of many cancers, but Song focuses on three main types: melanoma, gliomas, and estrogen receptor positive breast cancer.

Song’s recent developments have been published in a series of papers that analyze the functional consequences of non-coding genetic variants and mutations. Most genetic variants that modulate the risk to human diseases occur in these regions, which often regulate key gene expressions and remain difficult to understand. Some of these non-coding genetic variants have been discovered to increase the risk of developing cancer.

In particular, Song has been investigating the function of non-coding single nucleotide polymorphisms in the context of estrogen receptor positive breast cancers and low-grade gliomas.

“I often collaborate with Pablo Perez-Pinera, a member of the Cancer Center at Illinois [on these projects]. Our collaboration involves engineering predicted functional non-coding genetic variants and studying their molecular function. We will continue to work together to optimize genome editing techniques,” said Song.

Song has also collaborated with Paul Selvin, another member of the Cancer Center, to study pathogenic protein-DNA interactions at non-coding genetic variants that may contribute to low-grade gliomagenesis.

This research was funded in part by a Cancer Center at Illinois Planning Grant, awarded to Pablo Perez-Pinera, Jun Song, and Thomas Kuhlman, and National Cancer Institute R01CA163336.

Read Song’s series of papers studying the molecular function of non-coding genetic variants in cancer:

M. Manjunath†, J. Yan†, Y. Youn, K.L. Drucker, T.M. Kollmeyer, A.M. McKinney, V. Zazubovich, Y. Zhang, J.F. Costello, J. Eckel-Passow, P.R. Selvin, R.B. Jenkins, and J.S. Song (†co-first authors). Functional analysis of low-grade glioma genetic variants predicts key target genes and transcription factors, Neuro-Oncology, in press, (2020).

M. Majunath, Y. Zhang, S. Zhang, S. Roy, P. Perez-Pinera, J.S. Song. ABC-GWAS functional annotation of estrogen receptor-positive breast cancer genetic variants. Frontiers in Genetics, 11:730, (2020).

Y. Zhang, M. Manjunath, J. Yan, B.A. Baur, S. Zhang, S. Roy, J.S. Song. The cancer-associated genetic variant rs3903072 modulates immune cells in the tumor microenvironment. Frontiers in Genetics, 10:754, (2019).

Y. Zhang*, M. Manjunath*, S. Zhang, D. Chasman, S. Roy, and J.S. Song. (*co-first authors). Integrative genomic analysis predicts causative cis-regulatory mechanisms of the breast cancer-associated genetic variant rs4415084. Cancer Research, 78(7), 1579-1591, (2018).