Importance of scientific collaboration in contemporary drug discovery and development: a detailed network analysis

Abstract

Background: Growing evidence shows that scientific collaboration plays a crucial role in transformative innovation in the life sciences. For example, contemporary drug discovery and development reflects the work of teams of individuals from academic centers, the pharmaceutical industry, the regulatory science community, health care providers, and patients. However, public understanding of how collaborations between academia and industry catalyze novel target identification and first-in-class drug discovery is limited.

Results: We perform a comprehensive network analysis on a large scientific corpus of collaboration and citations (97,688 papers with 1,862,500 citations from 170 million scientific records) to quantify the success trajectory of innovative drug development. By focusing on four types of cardiovascular drugs, we demonstrate how knowledge flows between institutions to highlight the underlying contributions of many different institutions in the development of a new drug. We highlight how such network analysis could help to increase industrial and governmental support, and improve the efficiency or accelerate decision-making in drug discovery and development.

Conclusion: We demonstrate that network analysis of large public databases can identify and quantify investigator and institutional relationships in drug discovery and development. If broadly applied, this type of network analysis may help to enhance public understanding of and support for biomedical research, and could identify factors that facilitate decision-making in first-in-class drug discovery among academia, the pharmaceutical industry, and healthcare systems.

Keywords: Cardiovascular disease; Collaboration network; Drug discovery; Network analysis; PCSK9; Scientific collaboration; TNF inhibitors.

Fig. 1

Projecting paper institutions and references to the institutional collaboration network and the institutional knowledge flow network. a Paper I written by authors from institution a and b cite paper II written by authors from institution c, d and e, and paper III written by authors from institution c and d. b Collaborations among the five institutions based on the affiliations in the three papers. Link strength between institution c and d is 2; other link strengths are 1. c Directed links indicate the knowledge flows from institution c, d, and e to institution a and b; links from c/d to a/b have weight 2 and links from e to a/b have weight 1

Fig. 2

PCSK9 target discovery and development network analysis. a The number of publications and number of citations for PCSK9 papers by year. b Collaboration network in the discovery of PCSK9 for the top 20 institutions. Stripe width between institutions corresponds to the collaboration strength, i.e., the number of cases in which the two institutions collaborate. c The citation flow from cited papers (left) to citing papers (right). Stripe width from institutions on the left to institutions on the right corresponds to the number of cases in which papers from institutions on the left are cited by papers from institutions on the right

Fig. 3

Publication and citation growth. The number of annual publications (column a, c, e, g) and the number of annual citations (column b, d, f, h) for a, b 3 PCSK9 inhibitors (alirocumab, evolocumab, and bococizumab), c, d 3 PDE5 inhibitors (vardenafil, tadalafil, and sildenafil), e, f 8 HMG-CoA reductase inhibitors (cerivastatin, pitavastatin, fluvastatin, lovastatin, rosuvastatin, pravastatin, simvastatin, and atorvastatin,), and g, h 5 TNF inhibitors (certolizumab pegol, golimumab, etanercept, adalimumab, and Infliximab). In total, 170,099,684 publications dating from 1900 to 2017 were analyzed (see the “Methods” section)

Fig. 4

PCSK9 inhibitors network analysis. a–c Collaboration network for the top 20 institutions. Stripe width between institutions corresponds to the collaboration strength. d–f The citation flow for the top institutions. Stripe width from institutions on the left to institutions on the right corresponds to the number of cases in which papers from institutions on the left were cited by papers from institutions on the right