Cancer Mutation Analysis Through AI- A Strong Step Forward

When single-cell data is combined with a self-learning algorithm, it is a revelation as to how structural changes can go on to trigger cancer. This new methodology can indeed pave the road for cancer treatments that are personalised in nature.

Since cancer has many faces, it has a range of cancer mutations too. The totality of genomic alterations across an individual is what experts term the “mutational landscape.” Depending on the kind of cancer, these cancers differ from one another. It is well to be noted that people suffering from the same kind of cancer also have varied mutations.

Researchers have already documented mutational landscapes when it comes to different types of cancer. Somatic structural variants SVs account for more than half of the cancer-driving mutations. These happen to be those mutations in cells that pop up over the lifespan, like during cell division, when copying errors slide into the DNA and hence amend the structure of the chromosome.

These are hereditary in nature and are identified only in the cells that are affected and in the daughter cells. As age catches up, such genomic alterations tend to become more numerous, and the person’s mutation landscape resembles a unique mosaic.

Even though somatic SVs play a key role in the development of cancer, very little is known about them. As per Dr. Ashley Sanders, head of the Genome Stability and Somatic Mosaicism Lab at the Max Delbruck Center, there is still a dearth of methods that analyse effects on cell function. However, this is changing because of the new research findings that were recently published in one of the journals. A computational analysis method to identify as well as detect somatic EVs functional effect has been developed.

This has enabled the team to comprehend the molecular consequences related to individual somatic mutations across varied leukaemia patients, giving them new insights. It is also possible, according to Sanders, to use these findings to bring about therapies that target the mutated cells, thereby opening new gates when it comes to personalised medicine.

All these calculations happen to be based on Strand-seq data, which is an exclusive cell sequencing method that Sanders played a critical role in bringing about, and it was in 2012 that it was first introduced to the scientific community. This particular technique can analyse the cell genome in a much more detailed way than what happens with most conventional technologies.

Because of an experimental protocol which is sophisticated, this method can individually examine the two parental strands of the DNA.

Taking the work further, the research team is also able to gauge the nucleosome positions in each cell. Nucleosomes are DNA units that are wrapped around histone protein complexes. During gene expression, the position of the nucleosomes can differ with the wrapping, revealing if the gene is in an active state or not. The self-learning algorithm has been developed by Sanders which compares the gene activity within the patient cells either with or without SV mutations, thereby allowing to gauge the molecular impact of the structural variants. Sanders explains that they can now take a patient sample, have a look at the mutations that caused the disease, and also learn about the signalling pathways that these mutations which are disease causing disrupt. The team, for instance, was able to identify a unique but aggressive mutation in one of the leukaemia patients. This nucleosome analysis helped researchers with the data and specifics of the signalling pathways that are involved.

Sanders says that a single test can speak volumes about the mechanism involved related to cells in cancer formation and one can thereby use this knowledge to bring to fore customized treatments which are specific to every patient’s unique condition.