by Arturo Lopez Pineda
We are all encouraged to live a healthy lifestyle to avoid potentially life-threatening diseases. Exercising and good dietary habits make a big difference in maintaining our health. However, for some diseases, our cells carry important information that can alter this equation. It is estimated that the cells in our body have about 30 thousand genes. The information they encode tells each cell how to behave within our body. For example, a particular gene might determine the eye color of a person; another gene might tell a cell that it should become heart tissue; and yet another could be in charge of producing insulin in our body. However, sometimes these genes can be mutated, causing the gene to be either nonfunctional or functioning with a different behavior. These mutations have been found to cause some of the most challenging diseases.
Bioinformaticians are scientists that develop computational methods to help biologists analyze the vast quantity of information contained in our genes. Then, they come up with explanations or ideas on how gene mutations affect our health. They make intensive use of computers to create algorithms that determine whether a disease may develop. Consider Alzheimer’s disease, a degenerative disease of the brain that causes a gradual loss of memory. Bioinformaticians look for mutations in specific genes to determine if a person will develop it (learn more about Alzheimer’s disease). When these genes are altered, large amounts of toxic proteins are produced in the brain, causing the development of the disease.
There is one particular gene that is very famous among scientists, TP53. Referred to by bioinformaticians as the ‘guardian of the genome’, it has a role in preventing mutations and conserving stability in our DNA. TP53 is of great importance in preventing most diseases.
“Personalized Medicine” is a nascent field that tailors diagnosis and treatment to a patient by analyzing their clinical and genomic information. This is where bioinformaticians are assisting clinicians to achieve better diagnosis, treatments and clinical outcomes. Computer algorithms are a crucial part in this process, since human researchers cannot process the vast amount of information and interactions in the genomic data.
Algorithms can take into consideration a wide range of variables, including clinical signs and symptoms, laboratory data, and information from the DNA, such as the functioning of genes. They combine this information from a wide selection of people to come up with a model that can predict reasonably well the presence of a given disease. The rationale for this is to allow computers to ‘learn from past experiences’, and progressively gather data to improve upon their decisions.
As a vignette, consider a Latin American woman in her 50’s who notices a mass within her breast (for most Latin American countries, breast cancer is increasingly becoming a major burden of disease). In most cases she will present herself to her physician, where she will be tested by means of a mammogram (X-ray). The radiologist will then assess the image and a biopsy may be sought by a breast specialist if something suspicious is found. Once laboratory results are returned to her physician, she may, at that point be diagnosed with ‘breast cancer’. Using the information contained in genes, bioinformaticians have identified so far four distinct subtypes of breast cancer; Luminal A, Luminal B, HER2, and Basal-like, though it is very likely that further refinements of subtypes might be found. (For more information visit: Susan G. Komen Foundation). This is important, because the response to a given therapy will depend upon the subtype that the patient has within their genetic profile. A computer algorithm would ‘learn from past experiences’, meaning that it would identify the most important genes that have been found to explain the occurrence of that particular subset of cancer in patients. A suitable treatment regimen may be found for the patient combining the results from the algorithm with the expertise of the physicians. Fortunately, in the case of breast cancer, tests already exist that allow doctors to identify what breast cancer subtype the patient is suffering from – for other diseases, this is not yet the case.
In our laboratory we focus our efforts on the discovery of genes that could have some diagnostic role for breast cancer, lung cancer and brain tumors. We are particularly interested in finding those genes and their molecular characteristics that predict the subtype of cancer, the stage of cancer and associated suitable treatments. We create computer algorithms to facilitate this task and we continuously work to develop further refinements for these algorithms. In the future, as more genomic information is being made available, we will be able to integrate information from multiple sources and create richer testing platforms that can be easily accessible to patients and their doctors.
Arturo Lopez Pineda is a 2010 fellow of the Fulbright Science & Technology Award Program, from Mexico, and a PhD candidate in the Department of Biomedical Informatics at the University of Pittsburgh School of Medicine.