Use this URL to cite or link to this record in EThOS:
Title: Plant extracts and natural products : predictive structural and biodiversity-based analyses of uses, bioactivity, and 'research and development' potential
Author: Amirkia, V.
ISNI:       0000 0004 7231 0438
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
The process of drug discovery and development over the last 30 years has been increasingly shaped by formulaic approaches and natural products – integral to the drug discovery process and widely recognized as the most successful class of drug leads – have significantly been deprioritized by a struggling worldwide pharmaceutical industry. Alkaloids - historically the most important superclass of medically important secondary metabolites - have been used worldwide as a source of remedies to treat a wide variety of illnesses yet, there exists a wide discrepancy between their historical and modern significances. To understand these trends from an insider’s perspective, 52 senior-stakeholders in industry and academia were engaged to provide insights on a series of qualitative and quantitative aspects related to developments in the process of drug discovery from natural products. Stakeholders highlighted the dissonance between the perceived high potential of natural products as drug leads and overall industry and company level strategies. Many industry contacts were highly critical to prevalent company and industry-wide drug discovery strategies indicating a high level of dissatisfaction within the industry. One promising strategy which respondents highlighted was virtual screening which, to a large extent has not been explored in natural products research strategies. Furthermore, the physicochemical features of 27,783 alkaloids from the Dictionary of Natural Products were cross-referenced to pharmacologically significant and other metrics from various databases including the European Bioinformatics Institute’s ChEMBL and Global Biodiversity Information Facility’s GBIF biodiversity data. The combined dataset revealed that a compound's likelihood of medicinal use can be linked to its host species’ abundance and was input into target-independent machine learning algorithms to predict likelihood of pharmaceutical use. The neural network model demonstrated an accuracy of >57% for all pharmaceutical alkaloids and 98% of all alkaloids. This study is the first to incorporate the biodiversity of host organisms in a machine learning scheme characterizing druglikeness and thus demonstrates the link between host species’ abundance and druglikeness. These findings yield new insights into cost-effective, real-world indicators of drug development potential across the diverse field of natural products.
Supervisor: Heinrich, M. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available