Cambridge Cell Networks, a bioinformatics startup based in Cambridge, UK, is departing a bit from its peers in the computational systems biology market.
The company markets a database of around 500 hand-curated pathways, but rather than market it to a broad swathe of the biomedical research market, the company is targeting it to a niche user group — toxicologists.
In line with that strategy, CCNet has integrated its pathway database with around 20,000 chemical structures and their protein interactions in order to predict the toxic effects of compounds prior to animal experiments.
The resulting product, called ToxWiz, has been under development for around three years and required around 50 man-years of curation time, according to Gordana Apic, executive director of CCNet.
This week, the company announced that AstraZeneca has licensed ToxWiz for use in its toxicology programs. The firm is the first pharmaceutical customer that CCNet has disclosed, although Apic said that the company has signed other customers that it plans to announce in the next couple of weeks.
AstraZeneca actually played a key role in the development of ToxWiz, Apic said. “They have been giving us very valuable input on how to design the product properly in order to speed up drug development — specifically the tox issues,” she said. “They wanted something to bring together biology and chemistry and pathology and they asked us if we could use a systems biology framework to seamlessly get information together in a user-friendly way for people who don’t have much time.”
Scott Boyer, head of computational toxicology at AstraZeneca, said in a statement that ToxWiz “is one of the few software packages allowing integration of a range of complex data from a toxicology viewpoint.”
Boyer could not be reached for further comment before press time.
CCNet was founded in 2002 as a spinout from the University of Cambridge and has been profitable since 2004, Apic said. The company has additional sites in Heidelberg, Germany, and Serbia and employs around 30 staffers, 20 of whom are curators.
Apic said that ToxWiz elucidates relationships between proteins, genes, chemicals, pathologies, and toxic endpoints.
The pathway component of the resource, which includes links to around 800,000 journal abstracts, comprises only about a third of the offering, she said. “The distinguishing point that we have is about 20,000 chemical structures curated to their interactions with proteins, and to a level that we know which chemical structure interacts with which protein sequence.”
This enables users to search the database by chemical structure to find similar structures and then link to biological pathways in order to investigate the mechanism of action for the compound, or link to the likely toxic endpoints for the compound.
ToxWiz also includes a software component that enables users to predict whether a compound will be toxic or not — even in cases where the relationship between a chemical and an enzyme or metabolite is not known. “We use network context algorithms to predict … which new chemical will be metabolized by which enzyme,” she said.
As an example, Apic noted a case in which a clinician or toxicologist observes that a mouse treated with a certain compound has an enlarged liver. “They could just type in ‘liver hyperplasia’ or ‘enlarged liver’ and our system will perform the text search on internal data and find all the genes and proteins and chemicals associated with enlarged liver,” she said. “Then in parallel they can make a combined search with their chemical structure and the system will give them all the pathways, genes, and proteins associated with this chemical. So they can combine them and see how high the probability is that their chemical is causing liver hyperplasia, and what is the mechanism of it.”
“[AstraZeneca] wanted something to bring together biology and chemistry and pathology and they asked us if we could use a systems biology framework to seamlessly get information together in a user-friendly way for people who don’t have much time.”
Apic said that most of CCNet’s customers use ToxWiz in combination with experimental approaches “in order to reduce the number of animal studies.” By elucidating the mechanisms of toxicity, she said, researchers are able to design more efficient studies.
Other researchers are using ToxWiz as a reference database, she said, particularly in order to avoid time-consuming literature searches. “Literature searches normally come with a high number of false positives, and they’re often not linked to the right gene or protein or there is no obvious link,” she said. “That has been compiled for them in ToxWiz.”
Apic said that customers who have ToxWiz installed internally receive quarterly updates, while those who access the database through the company’s server receive weekly updates.
The company is currently planning to launch a new product targeted to clinicians. It will be based on a curated database approach similar to ToxWiz, but “is mostly to do with drug-drug interactions and even alternative indications for drugs in clinical development,” Apic said.
The new product will be released in the next couple of months, she said.