Distinction between Listeria species and L. monocytogenes serotypes with a handheld, VNIR hyperspectral imaging camera

Abstract A handheld, visible near-infrared (VNIR) hyperspectral imaging (HSI) camera was used to differentiate between Listeria species ( L. monocytogenes , L. ivanovii and L. innocua ) and two L. monocytogenes serotypes: 1/2a and 4b. The results indicated that portable HSI cameras could effectively be used for pathogen detection and distinction applications, despite lower spectral resolution and higher sensitivity to noise when compared to benchtop CCD-based instruments. Five hundred and fifty-five hyperspectral images were collected from pure cultures on agar plates throughout seven sampling campaigns. Sixty-five percent of samples/isolates ( n =± 90 samples per class) were used to train Partial Least Squares Discriminant Analysis (PLS-DA) models and 35% of samples ( n =± 50 samples per class) were used for validation. An overall classification accuracy (validation) of 74.5% was achieved for species-level discrimination through a four-class model ( L. monocytogenes 1/2a, L. monocytogenes 4b, L. ivanovii and L. innocua ), which demonstrates moderate performance within the VNIR range. However, Species-level performance was uneven across classes, with lower validation accuracies for L. innocua (68.0%) and L. ivanovii (63.0%) compared to L. monocytogenes 1/2a (71.8%) and 4b (95.2%). In contrast, a two-class model achieved ~ 90% accuracy for serotype-level discrimination between L. monocytogenes serotypes 1/2a and 4b. This highlights a novel use of portable VNIR cameras as a rapid, non-destructive serotyping aid. Variable Importance in Projection (VIP) score plots were used to characterise valuable wavelengths in the visible range that were significant for discriminant analysis, for example, 690 nm was significant for the classification of L. innocua , whilst 871 nm was associated with L. ivanovii. Notably, all the significant absorption regions used to differentiate between the L. monocytogenes serotypes were in the visible range (480–525 nm, 540–560 nm, 610 nm and 690 nm) despite the colony morphology being virtually identical for these organisms on brain heart infusion agar. Key points • A portable HSI camera was used to distinguish between Listeria colonies on agar • PLS-DA accuracy of 74.5% was obtained for classification between Listeria species • PLS-DA accuracy of 90% obtained for classification between L. monocytogenes serotypes

Authors

Institutions

Publication Details

Journal
Applied Microbiology and Biotechnology
Published
2026-10-26
DOI
https://doi.org/10.1007/s00253-026-14079-3
Primary Topic
Spectroscopy and Chemometric Analyses
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Distinction between Listeria species and L. monocytogenes serotypes with a handheld, VNIR hyperspectral imaging camera

Paul J. Williams, Diane Rip, Jana Schreuder
Applied Microbiology and Biotechnology
Spectroscopy and Chemometric Analyses
article

Distinction between Listeria species and L. monocytogenes serotypes with a handheld, VNIR hyperspectral imaging camera

Paul J. Williams, Diane Rip, Jana Schreuder
article en

Abstract

Abstract A handheld, visible near-infrared (VNIR) hyperspectral imaging (HSI) camera was used to differentiate between Listeria species ( L. monocytogenes , L. ivanovii and L. innocua ) and two L. monocytogenes serotypes: 1/2a and 4b. The results indicated that portable HSI cameras could effectively be used for pathogen detection and distinction applications, despite lower spectral resolution and higher sensitivity to noise when compared to benchtop CCD-based instruments. Five hundred and fifty-five hyperspectral images were collected from pure cultures on agar plates throughout seven sampling campaigns. Sixty-five percent of samples/isolates ( n =± 90 samples per class) were used to train Partial Least Squares Discriminant Analysis (PLS-DA) models and 35% of samples ( n =± 50 samples per class) were used for validation. An overall classification accuracy (validation) of 74.5% was achieved for species-level discrimination through a four-class model ( L. monocytogenes 1/2a, L. monocytogenes 4b, L. ivanovii and L. innocua ), which demonstrates moderate performance within the VNIR range. However, Species-level performance was uneven across classes, with lower validation accuracies for L. innocua (68.0%) and L. ivanovii (63.0%) compared to L. monocytogenes 1/2a (71.8%) and 4b (95.2%). In contrast, a two-class model achieved ~ 90% accuracy for serotype-level discrimination between L. monocytogenes serotypes 1/2a and 4b. This highlights a novel use of portable VNIR cameras as a rapid, non-destructive serotyping aid. Variable Importance in Projection (VIP) score plots were used to characterise valuable wavelengths in the visible range that were significant for discriminant analysis, for example, 690 nm was significant for the classification of L. innocua , whilst 871 nm was associated with L. ivanovii. Notably, all the significant absorption regions used to differentiate between the L. monocytogenes serotypes were in the visible range (480–525 nm, 540–560 nm, 610 nm and 690 nm) despite the colony morphology being virtually identical for these organisms on brain heart infusion agar. Key points • A portable HSI camera was used to distinguish between Listeria colonies on agar • PLS-DA accuracy of 74.5% was obtained for classification between Listeria species • PLS-DA accuracy of 90% obtained for classification between L. monocytogenes serotypes

Applied Microbiology and Biotechnology
Stellenbosch University (ZA)
Universiteit Stellenbosch
Good health and well-being
Openalex Percentile: Top 20%
Spectroscopy and Chemometric Analyses
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon