The human genome and drug discovery after a decade. Roads (still) not taken

Ruth Isserlin, Gary D. Bader, A.m. Edwards, Stephen V. Frye, Timothy M. Willson, Frank H. Yu

Abstract

The draft sequence of the human genome became available almost a decade ago but the encoded proteome is not being explored to its fullest. Our bibliometric analysis of several large protein families, including those known to be "druggable", reveals that, even today, most papers focus on proteins that were known prior to 2000. It is evident that one or more aspects of the biomedical research system severely limits the exploration of the proteins in the 'dark matter' of the proteome, despite unbiased genetic approaches that have pointed to their functional relevance. It is perhaps not surprising that relatively few genome-derived targets have led to approved drugs.

Introduction

The completion of the Human Genome Project was heralded as a turning point for medicine, promising a comprehensive "parts list" for human biology. However, the anticipated surge in drug discovery has been slow to materialize. This study investigates a systemic bottleneck known as the "Harlow-Knapp (H-K) effect"—the tendency of the biomedical community to concentrate research efforts on a small fraction of the proteome that was already well-characterized before the year 2000. We explore whether this stagnation is a result of institutional biases in funding and peer review, or a practical lack of research tools such as chemical probes and antibodies needed to explore the "dark matter" of the proteome.

Methods and Data

We conducted a large-scale bibliometric analysis tracking citation patterns for major protein families, including the human protein kinome (518 members), nuclear hormone receptors (NRs, 48 members), and ion channels. Using NCBI Entrez utilities and automated scripts, gene symbols and aliases were queried against PubMed to quantify citations across different eras (e.g., pre-2002 vs. post-2002). These trends were then correlated with the historical availability of high-quality, commercially accessible chemical probes to test the hypothesis that tools drive discovery more than disease links alone.

Results

The Persistence of Research Path-Dependency

Our analysis reveals that the H-K effect is a pervasive feature across protein families. Despite the availability of the full genome sequence, research remains highly concentrated. In the human protein kinome, for instance, 10% of kinases accounted for 84% of citations between 1950 and 2002. By 2009, this concentration only marginally decreased to 74%.

Distribution of citations across the human protein kinome.

As shown in the figure above, a vast "dark matter" exists within the proteome: approximately 60% of the kinome (roughly 300 kinases) received only 5% of total citations in 2009. This suggests that the majority of the genome's potential remains unexplored.

Chemical Probes as Catalysts for Discovery

A critical finding of this study is that research intensity is more strongly linked to the availability of high-quality reagents than to initial genetic disease associations. While a genetic discovery (e.g., BRAF in cancer) can spark interest, sustained research often requires accessible tools. This is most evident in the Nuclear Receptor family, where several "orphan" receptors saw massive spikes in citations only after the development of specific chemical probes.

Spikes in citations for orphan nuclear receptors following tool development.

As illustrated in the comparison between historical and 2009 citation data, eight orphan receptors (marked with asterisks) experienced significant citation spikes following the introduction of commercially available chemical probes. This demonstrates that even potentially "druggable" targets remain ignored by the broader community until the physical means to study them are widely accessible.

Discussion and Conclusion

The data suggest that the biomedical research enterprise is inherently conservative. Peer-review and funding mechanisms often penalize the exploration of unknown proteins, labeling such efforts as "fishing expeditions" while rewarding lower-risk research in established areas. While it is challenging to definitively disentangle causation from correlation, the temporal correlation—where the introduction of probes precedes citation spikes—strongly supports the hypothesis that tools drive discovery.

To unlock the potential of the human genome, the research enterprise must prioritize the unbiased generation and open distribution of high-quality reagents—including probes, antibodies, and purified proteins—for the entire proteome. Only by providing the tools to explore the unknown can we move beyond the "roads already taken" and accelerate the discovery of new therapeutic targets.

Summary updated

Abstract

The draft sequence of the human genome became available almost a decade ago but the encoded proteome is not being explored to its fullest. Our bibliometric analysis of several large protein families, including those known to be "druggable", reveals that, even today, most papers focus on proteins that were known prior to 2000. It is evident that one or more aspects of the biomedical research system severely limits the exploration of the proteins in the 'dark matter' of the proteome, despite unbiased genetic approaches that have pointed to their functional relevance. It is perhaps not surprising that relatively few genome-derived targets have led to approved drugs.
BibTeX
@inproceedings{isserlin2011human,
  author = {Ruth Isserlin and Gary D. Bader and A.M. Edwards and Stephen V. Frye and Timothy M. Willson and Frank H. Yu},
  title = {{The human genome and drug discovery after a decade. Roads (still) not taken}},
  booktitle = {Nature},
  year = {2011},
  doi = {10.1038/470163a},
}
Figure