A View from the Dark Side

In 1995, when I left a faculty position at the University of Pennsylvania to join the fledgling bioinformatics unit at what was then SmithKline Beecham Pharmaceuticals, the friendly jibes by academic colleagues about “going over to the Dark Side” were, I suspect, only half in jest. There weren't all that many of us practicing bioinformatics at that time, and there was genuine concern that a brain drain to industry might curtail the training of a new generation. Dire talk of “eating our seed corn” made its way into print [1–3]. Other misgivings grew out of the first wave of human genome-wide data then arriving in profusion: the short single-pass cDNA sequences called expressed sequence tags (ESTs) that bubbled up new gene identifications seemingly on a daily basis and that appeared to be suddenly short-circuiting the stately, hierarchical progression of the genome effort through ever-finer stages of mapping while breakthroughs in sequencing technology were labored after [4–6]. The data-management challenges arising from this heady sampling of the genome were making a strong impression, in both the public and private sectors, and the as-yet-unresolved (and highly charged) question of the patentability of genes led to a land rush on intellectual property [7–9]. At the same time a surge of startups with liberal venture-capital grubstakes only increased the demand for skilled gene prospectors [10,11]. All these developments fed concerns that a reckless commercialization of the human genome would be somehow unfairly turbocharged by lavish spending and pure computational horsepower in industry, with legions of apostate academics mining the data.