Dataset of nontuberculous mycobacteria (NTM) images on Elite agar medium

This is an image dataset of non-tuberculous mycobacteria (NTM) isolated from heater cooler units (HCUs) using the recently developed NTM Elite agar medium (bioMérieux, Marcy-l’Étoile, France). Non-tuberculous mycobacteria (NTM) are waterborne pathogens commonly found in building water systems where they are a primary concern to vulnerable patient populations and can cause severe disease. NTM Elite agar is a highly selective medium that includes four antimicrobial agents. Samples were concentrated through 0.45 µm mixed cellulose esters filters and filters were directly placed on the medium. Starting from the seventh day, cultures were examined weekly for seven weeks. Images of colonies confirmed as NTM were captured using an iPhone 11 dual camera. The collection of the image dataset was done in laboratory, at day-light and at different day of growing. Images represent the top view of NTM Elite agar plate.
All images are encoded in the standard JPG format and each image is named with an alphanumeric label whose first 4 letters define the species of the Mycobacterium, the number indicates the isolation number, followed by the day on which the photo was taken. For example, the label “MCHE1_4gg” indicates that the species in the photo is Mycobacterium chelonae, that is the isolation number 1 for this species and the photo was taken on day 4 of the incubation period. The list of mycobacteria species abbreviations is listed in Table 1.
To maintain an organized structure, the dataset is divided into two main sections: ORIGINAL and AUGMENTED.
The first main folder “ORIGINAL” contains the original images within ten subfolders, corresponding to the different species of NTM isolated, for a total of 92 images. The second main folder “AUGMENTED” contains the augmented images, also divided into ten subfolders. We performed some types of augmentation techniques such as hue (between −5° and +10°), saturation (between −10 % and +10 %), brightness (between −10 % and +10 %) and exposure adjustments (between −5 % and +5 %), to diversify and enrich the dataset, simulating also lighting changes.
The folders “Mycobacterium gordonae” and “Mycobacterium paragordonae” contain images of mycobacteria colonies that have a very similar appearance. These species are closely related; in fact the phylogenetic tree based on the 16S rRNA gene sequences of these two species shows a 99.0% similarity: it is therefore understandable that they are almost indistinguishable.
The images in the folder “More than one mycobacterial species” present more than one type of colonies of Mycobacteria per plate: arrows of different colors indicate colonies belonging to different species. Inside this folder an excel file, named “arrows and species” lists the correspondence between the colors of the arrows and the species for each image.
In the repository, there is also an Excel file that contains the list of images along with information related to the strain and the day of growth.

本数据集为非结核分枝杆菌（NTM）图像集，该菌种自加热冷却单元（HCUs）中分离得到，采用近期研发的NTM Elite琼脂培养基（bioMérieux，法国马西-莱图尔）进行分离。非结核分枝杆菌（NTM）是常见的水生病原体，常存在于建筑供水系统中，对易感人群构成主要威胁，并可引发严重疾病。NTM Elite琼脂是一种高度选择的培养基，包含四种抗菌剂。样品通过0.45 µm混合纤维素酯滤膜进行浓缩，滤膜直接置于培养基上。自第七天起，每周对培养物进行一次检查，共观察七周。使用iPhone 11双摄像头捕捉确认为NTM的菌落图像。图像数据集的收集在实验室进行，于日光照射下及培养的不同阶段完成。图像展示了NTM Elite琼脂平板的顶部视图。所有图像均采用标准的JPEG格式编码，每个图像均以字母数字标签命名，其中前四个字母定义了分枝杆菌的物种，数字表示分离编号，后跟照片拍摄的日期。例如，标签“MCHE1_4gg”表示图像中的物种为龟分枝杆菌，该物种的分离编号为1，照片是在培养期第4天拍摄的。分枝杆菌物种缩写列表见表1。为了保持组织结构的有序性，数据集分为两个主要部分：原始部分和增强部分。第一主要文件夹“原始”包含十个子文件夹中的原始图像，对应不同种类的NTM分离，共计92张图像。第二主要文件夹“增强”包含增强图像，同样分为十个子文件夹。我们对数据集进行了多种增强技术，如色调（-5°至+10°）、饱和度（-10%至+10%）、亮度（-10%至+10%）和曝光度调整（-5%至+5%），以丰富数据集，并模拟光照变化。文件夹“戈登分枝杆菌”和“旁戈登分枝杆菌”包含具有非常相似外观的分枝杆菌菌落图像。这两种物种亲缘关系密切；实际上，基于这两个物种的16S rRNA基因序列构建的进化树显示99.0%的相似性，因此它们几乎无法区分。文件夹“多种分枝杆菌物种”中的图像在同一平板上呈现多种分枝杆菌菌落：不同颜色的箭头指示属于不同物种的菌落。在此文件夹内，还包含一个名为“箭头与物种”的Excel文件，列出了每张图像中箭头颜色与物种之间的对应关系。在存储库中，还有一个Excel文件，包含图像列表以及与菌株和生长日期相关的信息。