Cancer is a highly complex and heterogeneous disease, where each tumor is unique. This creates challenges for managing the disease, as not all patients respond similarly to treatment. Therefore, researchers are putting effort into developing personalized treatment strategies for groups of patients with tumors that share certain characteristics.
To identify these characteristics, doctors and researchers from around the world have gathered massive amounts of data on the genetic and molecular make-up of tumors from different patients.
But to be able to decipher all this data, researchers must rely on computational approaches to find meaningful patterns that may point to personalized drug targets. In a recent publication, Tatiana Belova and colleagues present a computational method, named PORCUPINE, for analyzing such large, complex datasets.
– We have developed a computational method that can extract the important information from large datasets based on gene expression in different tumors. This allows us to identify subgroups of patients with tumors that share certain characteristics that can be used for developing personalized treatments, explains Belova.
Analysis of gene regulatory networks
Specifically, PORCUPINE is a computational tool that can analyze large datasets based on so-called gene regulatory networks. These networks contain information on the relationship between the genes within a tumor, and the proteins that control this expression, called transcription factors.
– By analyzing gene regulatory networks, we can find the molecular pathways that are involved in driving the differences we see between tumors from different patients. These pathways can then represent potential drug targets, says Belova.
– This way, PORCUPINE can be used to identify new potential drug targets or point to which existing treatment options may be most beneficial for each patient.
PORCUPINE helps identify potential drug targets in rare cancer
As part of their study, Belova and colleagues used PORCUPINE to analyze gene regulatory data from a rare and aggressive type of cancer that can form in smooth muscle tissue, called leiomyosarcoma.
Treatment of leiomyosarcoma is challenged by the high degree of heterogeneity between each case. Therefore, tailoring treatments for each patient based on their tumor characteristics, has the potential to improve the outcome of the disease.
– With the help of PORCUPINE, we were able to identify heterogeneously regulated pathways in leiomyosarcoma, that could represent potential drug targets for certain patient groups. We identified pathways that are known to promote leiomyosarcoma development, such as E2F signaling, as well as pathways that have been less well described in this cancer type, such as FGFR signaling. Both these pathways could be targets for personalized treatment, says Belova.
– Our goal is that this method can be used to stratify patients of any type of cancer, and pinpoint specific genes and pathways that could potentially serve as promising targets for cancer treatment, concludes Belova.
Publication
Belova, T., Biondi, N., Hsieh, P. H., Lutsik, P., Chudasama, P., Kuijjer, M. L. (2023). Heterogeneity in the gene regulatory landscape of leiomyosarcoma. NAR cancer, 5(3), zcad037. https://doi.org/10.1093/narcan/zcad037
Read more
PORCUPINE is available as an R package at GitHub.
An overview of the tools developed by the Kuijjer group can be found on their external website.