The Children’s Hospital of Philadelphia at night on Feb. 26, 2023.
Credit: Bamelak Duki
New artificial intelligence tools and datasets developed by researchers at the Wharton School and the Perelman School of Medicine aim to personalize treatments for kidney disease patients.
Katalin Susztak, director of the Penn/CHOP Kidney Innovation Center, and Wharton professor Nancy Zhang co-authored a paper discussing CellSpectra — “a computational tool that quantifies change in gene expression coordination across cellular functions” — applied to kidney and lung cancer data. The technologies can “analyze kidney disease at the cellular level to match the most effective treatments and speed up solutions,” according to Penn Today.
This atlas included “over 1 million cells, providing a comprehensive understanding of cell types, cell states and gene expression variations” and creates a plot called the single cell reference report card for the patient that points out the most disregulated cell type in the kidney, Susztak told The Daily Pennsylvanian.
“What we did here together with Nancy Zhang is we created a very large data set which we call SISKA, [which was] developed to map individual patient samples to this large reference atlas,” Susztak added.
SISKA stands for “Species Integrated Single cell Kidney Atlas.”
“We built CellSpectra to do what current methods cannot: analyze one patient’s sample at a time, and interpret it in the context of species, disease, and therapy,” Zhang told Penn Today.
There are over 850 million people who have a kidney disease worldwide, according to the International Society of Nephrology, with over a million dying of renal failure worldwide due to a lack of access to affordable treatment, according to the National Kidney Foundation.
New biomedical tools tools such as “single cell omics” or “single cell gene expression” allow for the characterization of every single cell in the kidney, Susztak said, adding that there are over 70 cell types just in the kidney.
“Using these large data sets, statistical methods and AI tools, we will be able to bring omics methods to the clinic,” she said. “It enables us to match patients with specific drugs that are already available, or we could generate new ones by targeting these genes and cell types.”
Susztak’s focus is kidney disease in the context of diabetes, which she said accounts for more than 50% of all kidney disease. It is “largest single cause of renal failure in the US and worldwide,” Susztak added.
Around ten years ago, Susztak championed the idea that more should be done for better patient characterization to enable new therapeutic developments for kidney disease.
“Way back, actually, this was not so fashionable, and I was not able to convince the National Institute of Health to fund the study,” she said.
Susztak turned to industry to fund potential drug development to that end, which she said she found greater success at Penn.
“I was very fortunate that Penn had a really good legal team who was able to put together pharma who is interested in diabetic kidney disease development with the academic centers… [to work] together in a non-competitive way,” Susztak said, adding that because pharma has “vested financial interest in direct development, [it was] very easy to convince everyone to work together in a non-competitive manner.”
“It shows quite a bit about pharma that they are able to put their immediate interest aside and are able to fund more discovery-based research,” Susztak added.
Susztak said that she would encourage medical researchers to work with pharma, but “every funding mechanism has its own pluses and minuses.”
Sign up for our newsletter
Get our newsletter, DP Daybreak, delivered to your inbox every weekday morning.
Susztak added that AI “has incredible potential” and is “slightly misunderstood.”
“Using AI tools the right way really enables people… to do the same things, essentially faster, more precisely, having information available quickly and reasonably well,” she said. “Our newest thing [is that] we have created the largest AI tool which essentially incorporates 400 terabytes of data just about the kidney.”
Susztak stressed that AI is not to replace doctors or pharma but to “enable the people who are already in the [field] to better diagnose [and] treat patients” alongside better matching patients “with [kidney disease] drugs and [to] identify new potential therapeutic targets.”
The Daily Pennsylvanian is an independent, student-run newspaper. Please consider making a donation to support the coverage that shapes the University. Your generosity ensures a future of strong journalism at Penn.