The world’s first fully machine-learning-powered breast cancer trial has just begun.
The researchers at Oxford University and Columbia University have identified how breast cancer cells are affected by mutations in their environment.
That means, with the help of AI, the scientists are now able to predict how long the patients will live, and the extent to which the mutations will lead to cancer.
It’s not just that the AI is getting smarter.
It’s also learning about the human body.
We know that a lot of cancer cells have mutations that have been inherited from the parents.
We’ve known that mutations in genes are involved in cancer, but until now, there has been very little information about the biology of how mutations affect the body.
The Oxford team has done all this.
“The first step is to identify a genetic sequence that is relevant to cancer,” says Dr. James Farr, an Oxford professor of AI and an author of the paper.
That genetic sequence is then used to build a model of a patient’s genome.
It tells the machine how many genes and proteins the patient has and how many different types of cells are present.
The model then learns how the patient reacts to different types and combinations of those genes and the proteins.
The machine also learns the genes and what the proteins do.
“This is an important step for the human breast, because the mutations are not necessarily reversible, and therefore it’s difficult to predict the response of a breast cancer patient to treatment,” says Farr.
“We’ve shown that we can use the model to predict what will happen to the cells of the patient and the tumour.”
For example, a mutation in the genetic code of the breast cancer-causing protein ERK1A, which is involved in the growth of tumors, could have a positive effect on the development of tumors and make them more aggressive, and it might be linked to a higher rate of cancer deaths.
That could mean the mutation might slow the rate of growth and development of cancer, the researchers say.
Other types of mutations in the body also affect the rate at which the body can repair itself, so it may have a negative effect.
“These changes may have important clinical implications,” says the researchers.
One of the biggest challenges for breast cancer researchers is figuring out how to predict exactly how a mutation will affect the patient, says Dr: Anwar Khan, a professor of medical genetics at Harvard Medical School and a member of the team.
“The only way to do that is to build models of the individual cells and how they respond to their environment,” he says.
“That’s a huge problem in cancer research because we can’t predict which mutation will cause a disease or which mutation can improve the outcome.”
A new approachThe researchers’ model of breast cancer is the first step in a larger research effort.
They have been working with a team of scientists from the University of Michigan, the University at Buffalo, the Rockefeller University and the University in Barcelona.
“It’s going to be important to develop models of different cancers for years to come, and then hopefully to build new models,” says Khan.
“Because the models are the best way to understand what’s going on in a patient, they’re also the most accurate.”
The researchers have been developing their model by combining several other approaches.
“Many cancer studies are based on the same cell lines, which are hard to model, so this model was developed in a way that is very different from previous models,” Khan says.
The model is also very much in line with the current state of breast tissue.
“Our goal was to model cells in their natural state, with no mutations,” says Kazi.
The scientists used the same model for breast tissue in mice, which has the advantages of being able to use more tissue than with mammary tissue, because it’s easier to observe changes in cell growth.
But it also makes it harder to model changes that might happen in a tumor.
The researchers have also been able to get rid of a number of genetic differences in breast tissue, such as differences in how much melanocytes, the white blood cells that help the body fight infections, are expressed.
In addition to the model, the team is working on a paper that describes a new approach for predicting the response to chemotherapy.
“What we want to do is predict the changes that occur in the cancer cells, how they grow and when they’re dying,” Khan explains.
“To do that, we need to build out models of all the different cells, which means learning how many types of cancer it has, and how those different cells respond to different chemotherapy drugs.”
The work is part of a larger effort by the Oxford team to build machines that can predict human disease, from how many years someone will live to how long they will live in different age groups.
In that sense, the Oxford research is a bit like a research laboratory.
The machine is able to see what’s happening in the human brain