A study on identification of bacteria in environmental samples using single-cell Raman spectroscopy: feasibility and reference libraries.

We report on our recent efforts towards identifying bacteria in environmental samples by means of Raman spectroscopy. We established a database of Raman spectra from bacteria submitted to various environmental conditions. This dataset was used to verify that Raman typing is possible from measurements performed in non-ideal conditions. Starting from the same dataset, we then varied the phenotype and matrix diversity content included in the reference library used to train the statistical model. The results show that it is possible to obtain models with an extended coverage of spectral variabilities, compared to environment-specific models trained on spectra from a restricted set of conditions. Broad coverage models are desirable for environmental samples since the exact conditions of the bacteria cannot be controlled.

High-resolution 3D phase imaging using a partitioned detection aperture: a wave-optic analysis.

Quantitative phase imaging has become a topic of considerable interest in the microscopy community. We have recently described one such technique based on the use of a partitioned detection aperture, which can be operated in a single shot with an extended source [Opt. Lett.37, 4062 (2012)OPLEDP0146-959210.1364/OL.37.004062]. We follow up on this work by providing a rigorous theory of our technique using paraxial wave optics, where we derive fully 3D spread functions for both phase and intensity. Using these functions, we discuss methods of phase reconstruction for in- and out-of-focus samples, insensitive to weak attenuations of light. Our approach provides a strategy for detection-limited lateral resolution with an extended depth of field and is applicable to imaging smooth and rough samples.

Single bacteria identification by Raman spectroscopy.

We report on rapid identification of single bacteria using a low-cost, compact, Raman spectroscope. We demonstrate that a 60-s procedure is sufficient to acquire a comprehensive Raman spectrum in the range of 600 to 3300 cm⁻¹. This time includes localization of small bacteria aggregates, alignment on a single individual, and spontaneous Raman scattering signal collection. Fast localization of small bacteria aggregates, typically composed of less than a dozen individuals, is achieved by lensfree imaging over a large field of view of 24 mm². The lensfree image also allows precise alignment of a single bacteria with the probing beam without the need for a standard microscope. Raman scattered light from a 34-mW continuous laser at 532 nm was fed to a customized spectrometer (prototype Tornado Spectral Systems). Owing to the high light throughput of this spectrometer, integration times as low as 10 s were found acceptable. We have recorded a total of 1200 spectra over seven bacterial species. Using this database and an optimized preprocessing, classification rates of ~90% were obtained. The speed and sensitivity of our Raman spectrometer pave the way for high-throughput and nondestructive real-time bacteria identification assays. This compact and low-cost technology can benefit biomedical, clinical diagnostic, and environmental applications.

Sparsity-driven reconstruction for FDOT with anatomical priors.

In this paper we propose a method based on (2, 1)-mixed-norm penalization for incorporating a structural prior in FDOT image reconstruction. The effect of (2, 1)-mixed-norm penalization is twofold: first, a sparsifying effect which isolates few anatomical regions where the fluorescent probe has accumulated, and second, a regularization effect inside the selected anatomical regions. After formulating the reconstruction in a variational framework, we analyze the resulting optimization problem and derive a practical numerical method tailored to (2, 1)-mixed-norm regularization. The proposed method includes as particular cases other sparsity promoting regularization methods such as l(1)-norm penalization and total variation penalization. Results on synthetic and experimental data are presented.

An efficient numerical method for general L(p) regularization in fluorescence molecular tomography.

Reconstruction algorithms for fluorescence tomography have to address two crucial issues: 1) the ill-posedness of the reconstruction problem, 2) the large scale of numerical problems arising from imaging of 3-D samples. Our contribution is the design and implementation of a reconstruction algorithm that incorporates general Lp regularization (p ¿ 1). The originality of this work lies in the application of general Lp constraints to fluorescence tomography, combined with an efficient matrix-free strategy that enables the algorithm to deal with large reconstruction problems at reduced memory and computational costs. In the experimental part, we specialize the application of the algorithm to the case of sparsity promoting constraints (L (1)). We validate the adequacy of L (1) regularization for the investigation of phenomena that are well described by a sparse model, using data acquired during phantom experiments.