Dm2500 microscope

The dynamic light scattering technique (DLS; 90 Plus, Brookhaven, NY, USA) was used to determine the mean droplet size, polydispersity index (PDI) and Zeta potential of the IL-based formulations at 25 °C, in triplicate. The measurement angles were 90° for particle size and 15° for Zeta potential. Optical microscopy was performed on a DM2500 microscope (Leica Microsystems, Wetzlar, Germany), equipped with a Moticam 480 camera (Motic, Barcelona, Spain). Transmission electron microscopy (TEM, high-resolution JEOL 300 kV) was performed via IL negative staining with 1% phosphotungstic acid [17 (link)].

Fluorescence images were acquired with a Leica-Microsystems DM2500 microscope. Optical microscope images were obtained with a Zeiss Axioplan 2 microscope, equipped with a MediaCybernetics Evolution VF digital camera. For images of the shape change this microscope was equipped with a Linkam LTS420 thermal stage under nitrogen atmosphere, using heating rates ranging from 3 to 10 °C min⁻¹. The movie of the “kicking” glass bead was taken with a Leica WILD M10 microscope, equipped with an Evolution MP5.0 Leica DMC2900 digital camera with frame rate of 30 fps and a Linkam LTS420 thermal stage under nitrogen atmosphere.

The smears were prepared according to the previously published protocol [36 (link)] and stained with either Giemsa or 4',6-diamidino-2-phenylindole (DAPI) as described before [37 , 38 (link)]. Digital images were acquired in DM 2500 microscope (Leica Microsystems GmbH, Wetzlar, Germany) equipped with UCMOS14000KPA 14-Mpx camera (Toup Tek, Hangzhou, China) at ×1,000 magnification. Giemsa-stained smears were observed and photographed in bright field (BF), living cells using differential interference contrast (DIC), whereas DAPI-stained smears with both DIC and fluorescence microscopy. All measurements of cells (n = 31) and statistical analysis were performed in UTHSCSA Image Tool for Windows v. 3.0.

To improve the model’s performance by increasing the quantity and diversity of images, we merged two data sources: Google images and microscopic images produced in our laboratory.
Initially, we searched for fungal species on Google images and downloaded publicly available images. Next, we took microscopic images of fungal strains prepared in our lab by obtaining slides from the 310 fungal isolates collected from Kaohsiung Veterans General Hospital from 2020 to 2021. The fungal isolates were grown on potato dextrose agar plates and incubated at 25°C. Microscopic images were taken from well-grown colonies using touch-tape preparation and lactophenol cotton blue (LPCB) staining. The images were captured using a DM2500 microscope (Leica Microsystems, Germany) with 40-fold magnification and an EOS 60D camera (Canon, Japan). We recorded 10 to 20 photos per slide with a characteristic morphology, and one slide was taken for each sample. Finally, two experienced medical technologists interpreted all image labels included in this study.

For mouse airways, immunofluorescence for RBM3 was performed on naïve and Alternaria challenged airways as previously reported^{64 (link),65 (link)}. Briefly, lung samples were de-paraffinized by sequential placement in xylene and ethanol. Staining for RBM3 was performed with rabbit polyclonal antibody (PeproTech) at 1:1000 concentration. Tyramide Signal Amplification Kit #41 (Invitrogen) was used for fluorescent signal amplification with subsequent DAPI staining (Vector Laboratories). Lung airways were visualized with a DM2500 microscope (Leica Microsystems). Cell nuclei were stained with DAPI. Images were taken from at least 5 airways of at least 3 mice per group.

All slides stained for histology and immunohistochemistry were acquired using a DM2500 microscope (Leica Microsystems, Wetzlar, Germany) equipped with a DFC310 FX digital CCD camera (Leica Microsystems). ImageJ software was used for the statistical analysis of level of bone formation on the designated regions of MTC staining tissues as described previously (Adhikari et al., 2019 (link)). Briefly, the original MTC stained images were converted to RGB images and they were deconvolved by ImageJ using color deconvolution plugin. When a constant threshold was set, the area and integrated density of collagen fibres in the region of interest were estimated. Four different images of 50 μm² from 5 different samples from randomly selected specimens were used. For the quantitative analysis of RUNX2 and OC, positive cells of them were counted from 3 images of 100 μm² (positive for immunostaining) selected randomly from 5 different specimens.

All slides prepared for histology and immunohistochemistry were examined using a DM2500 microscope (Leica Microsystems, Wetzlar, Germany) equipped with a DFC310 FX digital CCD camera (Leica Microsystems). ImageJ software was utilized for the statistical analysis of bone formation levels in the specified regions of Masson’s trichrome (MTC) stained tissues, following established protocol (Adhikari et al., 2019 (link)). In brief, the original MTC stained images were converted to RGB format and then deconvolved using the color deconvolution plugin in ImageJ. Subsequently, after setting a consistent threshold, the area and integrated density of collagen fibers within the region of interest were quantified. For this analysis, six different images covering an area of 200 μm² were analyzed from five randomly selected specimens. Additionally, the quantification of RUNX2 expression involved counting the number of positive cells in immunostained sections. This was performed using six images covering an area of 100 μm² each, obtained from five different specimens.