Project staff and support

Zofia Jones (Intern, University of Nottingham)
Gordon James and Adrian Smith (Company Supervisors, Unilever)
John King (Academic Mentor, University of Nottingham)
Lorcan Mac Manus (Technology Translator, Industrial Mathematics KTN)

This Internship project is being carried out at Unilever Ltd, in conjunction with University of Nottingham. It is part of the KTN's Industrial Mathematics Internships Programme, co-funded by EPSRC. Start date: April 2010; duration: 5 months.

Project description

The internship project will primarily encompass the construction of complex mathematical algorithms to aid in the development of an in silico kinetic model of axillary (underarm) malodour. Specifically, the student will be asked to address two significant deficiencies in Unilever’s current prototype model:

• Most importantly, we require an algorithm that can relate traditional culture-based microbiological data with modern microbiomics-type data. Currently, the data in the model, from lab-based kinetic studies, is expressed as molar levels of each odour molecule generated per million bacteria per hour. This can easily be related to data obtained via the traditional route by which microbes are cultured and enumerated from human skin, expressed as numbers of bacteria per unit area. However, modern data capture is radically different in that bacteria are not cultured from skin, but are instead lysed, and have one of their genes amplified and read by high-throughput DNA sequencing. Each bacterial type has a unique gene signature, and the data are expressed as numbers of each individual gene sequence per sample. Each data-point is influenced by the efficiency of bacterial lysis and gene amplification, and the numbers of signature genes carried by that particular bacterial type. By factoring these in, we wish to develop an algorithm that can interconvert such data to absolute bacterial numbers.
• Additionally, we require an algorithm that can relate aroma threshold levels for odour molecules with clinical malodour scores. Currently, the model contains objective data on the concentration of each odour molecule that is detectable by the human nose as malodorous. However, we need to build an algorithm to interconvert this data with subjective malodour scores, as determined by expert underarm sniff-assessors.

related resources:

»	Underarm malodour modelling
	Mathematical tasks

other projects:

[Find other Medical and pharmaceutical projects]