Interpreting support vector machines applied in laser-induced breakdown spectroscopy.

Laser-induced breakdown spectroscopy is often combined with a multivariate black box model-such as support vector machines (SVMs)-to obtain desirable quantitative or qualitative results. This approach carries obvious risks when practiced in high-stakes applications. Moreover, the lack of understanding of a black-box model limits the user's ability to fine-tune the model. Thus, here we present four approaches to interpret SVMs through investigating which features the models consider important in the classification task of 19 algal and cyanobacterial species. The four feature importance metrics are compared with popular approaches to feature selection for optimal SVM performance. We report that the distinct feature importance metrics yield complementary and often comparable information. In addition, we identify our SVM model's bias towards features with a large variance, even though these features exhibit a significant overlap between classes. We also show that the linear and radial basis kernel SVMs weight the same features to the same degree.

Laser-induced breakdown spectroscopy as a straightforward bioimaging tool for plant biologists; the case study for assessment of photon-upconversion nanoparticles in Brassica oleracea L. plant.

The main purpose of this work is to thoroughly describe the implementation protocol of laser-induced breakdown spectroscopy (LIBS) method in the plant analysis. Numerous feasibility studies and recent progress in instrumentation and trends in chemical analysis make LIBS an established method in plant bioimaging. In this work, we present an easy and straightforward phytotoxicity case study with a focus on LIBS method. We intend to demonstrate in detail how to manipulate with plants after exposures and how to prepare them for analyses. Moreover, we aim to achieve 2D maps of spatial element distribution with a good resolution without any loss of sensitivity. The benefits of rapid, low-cost bioimaging are highlighted. In this study, cabbage (Brassica oleracea L.) was treated with an aqueous dispersion of photon-upconversion nanoparticles (NaYF4 doped with Yb3+ and Tm3+ coated with carboxylated silica shell) in a hydroponic short-term toxicity test. After a 72-hour plant exposure, several macroscopic toxicity end-points were monitored. The translocation of Y, Yb, and Tm across the whole plant was set by employing LIBS with a lateral resolution 100 µm. The LIBS maps of rare-earth elements in B.oleracea plant grown with 50 µg/mL nanoparticle-treated and ion-treated exposures showed the root as the main storage, while the transfer via stem into leaves was minimal. On the contrary, the LIBS maps of plants exposed to the 500 µg/mL nanoparticle-treated and ion-treated uncover slightly different trends, nanoparticles as well as ions were transferred through the stem into leaves. However, the main storage organ was a root as well.

Detail investigation of toxicity, bioaccumulation, and translocation of Cd-based quantum dots and Cd salt in white mustard.

In this study, a model crop plant white mustard (Sinapis alba L.) was treated with an aqueous dispersion of silica-coated CdTe quantum dots (CdTe/SiO2 QDs) in a 72-h short-term toxicity test. The toxicity was established via measurements of (i) the root length and (ii) the chlorophyll fluorescence. These results were compared to two other sources of cadmium, free Cd ions (CdCl2) and prime un-shell nanoparticles CdTe QDs. Tested compounds were applied in concentrations representing 20 and 200 µM Cd. The uptake and translocation of Cd were investigated using inductively coupled plasma optical emission spectrometry (ICP-OES) and the spatial Cd distribution was investigated in detail applying laser induced breakdown spectroscopy (LIBS). The LIBS maps with a lateral resolution of 100 µm were constructed for the whole plants, and maps with a lateral resolution of 25 µm (micro-LIBS arrangement) were used to analyse only the most interesting parts of plants with Cd presence (e.g. root tips or a part crossing the root into the above-ground part). Our results show that the bioaccumulation patterns and spatial distribution of Cd in CdTe/SiO2 QDs-treated plants differ from the plants of positive control and CdTe QDs. Fluorescence microscopy photographs revealed that CdTe/SiO2 became adsorbed onto the plant surface in comparison to CdTe QDs. Further, a physico-chemical characterization of QDs before and after the test exposure showed only minor changes in the nanoparticle diameters and no tendencies of QDs for agglomeration or aggregation during the exposure.

Benchmark classification dataset for laser-induced breakdown spectroscopy.

In this work, we present an extensive dataset of laser-induced breakdown spectroscopy (LIBS) spectra for the pre-training and evaluation of LIBS classification models. LIBS is a well-established spectroscopic method for in-situ and industrial applications, where LIBS is primarily applied for clustering and classification tasks. As such, our dataset is aimed at helping with the development and testing of classification and clustering methodologies. Moreover, the dataset could be used to pre-train classification models for applications where the amount of available data is limited. The dataset consists of LIBS spectra of 138 soil samples belonging to 12 distinct classes. The spectra were acquired with a state-of-the-art LIBS system. Lastly, the composition of each sample is also provided, including estimated uncertainties.