Genome sequencing and beyond

At an afternoon session titled "Genome Analyses and Sequencing to Advance Drug Discovery and Treatment" researchers discussed how they are use sequencing tools to reveal new mutations that may underlie cancer.

Dr. Rick Wilson of Washington University described whole genome sequencing as a way to identify genes involved in cancer. He said that this technique is yielding tantalizing clues about cancer development. Each person's cancer is a little different, but with a large collection of patient samples, he and his colleagues hope to expand the list of candidate genes.

A great example of genetic variability within one category of cancer is a mutation in the gene EGFR in cancer. Mutations to this gene determine whether a patient will respond to the drug Iressa. Iressa was nearly dismissed because it had an effect on such a minority of patients until this genetic quirk was discovered: it turned out that 80% of the patients who responded to the drug had this mutation. Testing for EGFR can help doctors determine if this is the right drug for an individual.

Cancer researchers have a long list of candidate genes already associated with cancer, but what lies outside of this list? Using whole genome sequencing could be a way to find out. Dr. Wilson mentioned a sub-group of the genome he referred to as the "conservome" - genes are conserved across species, so researchers know that they must have some sort of important function...even if they don't know what that function is yet.

Dr. Wilson focused his talk on AML, a relatively common form of cancer with a particularly low survival rate. Sequencing the whole genome of one sample yielded 10 promising mutations, eight of which were "private" mutations, meaning that they had not been detected in previous samples. This is the promise of whole genome sequencing - the detection of mutations we would otherwise never look for.

Of course, many tumors are a mixture of cancer cells, and genome sequencing can only provide a glimpse of the genome at one point in time and in one location. The cost is decreasing, but it's still an expensive endeavor.

Dr. Wilson looked at samples from patients who had relapsed and discovered additional mutations, in part because the technology had improved, but also because new mutations had cropped up. He plans to sequence 50 genomes from 50 cases.

In addition to looking across a single cancer genome, it seems like a great idea to look across cancer types for mutations that transcend the categories of cancer we usually think about. For instance, Dr. Wilson mentioned a mutation in glioblastoma that cropped up in the list of new mutations in AML - unexpectedly, it was associated with good prognosis in GBM and poor prognosis in AML. Which left me thinking...what?! Seems like for the moment, whole genome sequencing is going to give us more questions to chew on than answers to satiate our curiosity.

The next speaker, Dr. Mary Relling from St. Jude Children's Hospital, addressed a provocative topic in genomics - race. Some drugs have a different effect in patients of different races (for instance, Plavix). There are also disparities in patient outcome that appear to vary based on race. Of course, the word race is a bit of a misleading term; if genomics tells us anything about race, it is that things are far more complex than the checkboxes we are supposed to choose from on forms that invite us to declare our singular race. Dr. Relling notes that she hopes we'll be able to move beyond self-declared race, using SNPs associated with ancestry to help determine the best drugs and treatment options for individual patients. Right now, we don't really need to type many SNPs to find ancestry. She noted that in her research, they really don't pay much attention to self-declared race and would like to get away from these self-imposed labels as quickly as possible, using genomics to make more informed decisions.

Dr. Relling also touched upon genetic diversity, reminding us that genetic diversity varies with geographic distance "as the person walks out of East Africa" rather than as the crow flies and that race is truly continuous as opposed to categorical.

Dan Roden of Vanderbilt University spoke next. Dr. Roden showed a slide of patient's medical record (with name and all identifiable pieces of information blurred out, for obvious reasons!). He described his dream of eventually being able to call up a list of all genes for a patient associated with a drug that a doctor might be considering administering. He described incorporating personal genomics into medical records. Vanderbilt has developed a way of "scrubbing" all identifying information from a patient's records for research purposes using techniques originally developed by the CIA. As doctors add information to the record, the anonymous version seen by researchers will be updated without allowing the researchers to access any information that can identify the patient.

I'm reading The Immortal Life of Henrietta Lacks right now, and was struck by the way that researchers today are going to painstaking efforts to protect patients' identities when not that many decades ago, a patient's life or the lives of a patient's family could be so utterly destroyed when a patient's identity was revealed. I'm glad privacy measures have finally caught up to Dr. Roden just needs that list of genes whose function in relationship to drugs is known and vetted.