A strong intuition is much more powerful than a weal test. Normals teach us rules; outliers teach us laws. For every perfect medical experiment, there is a perfect human bias.

If you take 100 breast-cancer samples, 100 types of cancer have 100 different hallmarks of mutated genes. You could be nihilistic and say, 'Oh, God, we'll never be able to tackle this!' But there are deep, systematic, organizational principles at work in all that diversity.

Because I work on leukemia, the image of cancer I carry in my mind is that of blood. I imagine that doctors who work on breast cancer or pancreatic cancer have very different visualizations.

Nearly every one of the genes that turns out to be a key player in cancer has a vital role in the normal physiology of an organism. The genes that enable our brains and blood cells to develop are implicated in cancer.

There is an enormous amount of options that a physician can provide today, right down from curing patients, treating patients, or providing patients with psychic solace or pain relief. So, in fact, the gamut of medical intervention is enormous.

Some cancers are curable, while others are highly incurable. The spectrum is enormous. Metastatic pancreatic cancer is a highly incurable disease, whereas some leukemia forms are very curable. There is a big difference between one form and another.

Science is among the most profoundly human of our activities. Far from being subsumed by the dehumanising effects of technology, science, in fact, remains our last stand against it.

My memory of my household is of one immersed in books and music. I have a very intimate relationship with Bengali literature, particularly Tagore, and my interest besides reading then was music.

Unlike other diseases, the vulnerability to cancer lies in ourselves. We always thought of disease as exogenous, but research into cancer has turned that idea on its head - as long as we live, grow, age, there will be cancer.

We may have to learn to live with cancer rather than die of it. It means a big change in our mindset and how we do research. We haven't quite reached there yet.