By Rebecca Milec
It may seem like every patient with adenocarcinoma or acute lymphoblastic leukemia is the same, but new technology developed by researchers at The Hospital for Sick Children (SickKids) proves that’s not the case. A new study published Sunday in Nature Methods introduces a new computational tool that can successfully take multiple types of data from patients and show the similarities and differences between these “similar” patients with the same disease or disorder.
This new method, called Similarity Network Fusion (SNF), is unique in that it is the first to introduce patient similarity networks (groupings of similar patients) and use them to effectively aggregate multiple forms of data, such as DNA, metabolomics, genomics and clinical information.
“SNF allows us to get a really clear picture into the similarities and differences between patients,” says Dr. Anna Goldenberg, Scientist in Genetics &Genome Biology at SickKids and Assistant Professor in the Department of Computer Science at the University of Toronto, whose lab developed this novel approach. “Recently, more similar disease subtypes have been discovered in many complex diseases such as cancer, but they are still very broadly classified. The reality is much more complex than that and patient networks can help us dissect the disease differences further than ever before.”
Tools like SNF could lead to a much more refined representation and understanding of diseases, phenotypes and other biological phenomena, says Goldenberg. This knowledge could eventually contribute to the development of individualized treatments for complex diseases and disorders, with an ultimate goal of applying this new methodology in the clinical setting. SNF could enable clinicians to see where their new patients fall in the spectrum of previous patients and use their patients’ data to provide a more precise diagnosis and prognosis, improving individualized care.
While this computational tool was initially developed to study data from patients of all ages in a clinical research setting, the possible applications are broad. The tool can be used in many different areas of research, even those that don’t involve humans, such as studying tomato strains for similarities in desired traits, like sweetness.
For more information on the tool and for the freely available SNF code please see