Development of a computer vision pipeline for the analysis of Aliivibrio fischeri bioluminescence inhibition on solid media

Date
2023-11
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH ABSTRACT: The South African Government aims to promote a circular economy, particularly upcycling organic waste, to improve soil quality in the Western Cape. Notably, wastewater sludge, abundant with organic nutrients, also contains harmful micropollutants. Thus, a reliable, on-site toxicity measurement method is essential. This study introduces an in-field technology to assess micropollutant toxicity in soil and water leachate samples and gauge bioremediation or bioaccumulation levels post-sludge application. The study leverages the bioluminescent properties of Aliivibrio fischeri, commonly known as Vibrio fischeri, as a toxicity indicator. Initial experiments using an optical microbial biosensor in liquid media showed inconsistent results due to hypoxic culture conditions. Subsequent experiments on solid media, employing computer vision and machine learning, demonstrated the consistent inhibition of Vibrio fischeri bioluminescence, serving as an effective toxicity marker. Test compounds, zinc sulphate and atrazine, validated the approach, aligning with existing literature on their toxicity levels. Machine learning models predicted treatment states and concentrations accurately using EC50 estimates. This research is foundational for a mobile diagnostic application for farmers to swiftly and accessibly gauge soil and water quality.
AFRIKAANSE OPSOMMING: Die Suid-Afrikaanse regering poog om ’n sirkulˆere ekonomie te bevorder, veral om organiese afval te hersirkuleer, om sodoende grondgehalte in die Wes-Kaap te verbeter. Veral afvalwaterslyk, wat vol organiese voedingstowwe is, bevat ook skadelike mikrobesoedeling. ’n Betroubare metode vir toksisiteitmeting ter plaatse is dus ’n noodsaaklikheid. Hierdie studie stel ’n in-veld-tegnologie bekend om mikrobesoedelende toksisiteit in grond- en waterloogmonsters te bepaal en bioremedi¨ering, of bioakkumulasievlakke na-slyktoediening, te meet. Die studie maak gebruik van die bioluminiserende eienskappe van Aliivibrio fischeri of Vibrio fischeri as ’n toksisiteitsaanwyser. Aanvanklike eksperimente met behulp van ’n optiese mikrobiese biosensor in vloeibare media het inkonsekwente resultate getoon as gevolg van hipoksiese kultuurtoestande. Daaropvolgende eksperimente op vaste media, met behulp van rekenaarvisie en masjienleer, het die konsekwente inhibisie van V. fischeri bioluminisering getoon, wat as ’n effektiewe toksisiteitsmerker dien. Toetsverbindings, sinksulfaat en atrasien, het die benadering bekragtig, in ooreenstemming met bestaande literatuur oor hul toksisiteitsvlakke. Masjienleermodelle het behandelingstoestande en konsentrasies akkuraat voorspel deur EC50-skattings te gebruik. Hierdie navorsing lˆe die grondslag vir ’n mobiele diagnostiese toepassing vir boere om grond- en watergehalte vinnig en toeganklik te meet.
Description
Thesis (MSc(Eng))--Stellenbosch University, 2023.
Keywords
Citation