Progress 01/15/23 to 01/14/24
Outputs
Target Audience:Poultry farms and processing plants. Changes/Problems:NA What opportunities for training and professional development has the project provided?My students presented the work at the AICHE meeting. How have the results been disseminated to communities of interest?Two conference presentations and one journal publication (impact factor: 3.5) What do you plan to do during the next reporting period to accomplish the goals?We plan to study the specificity of the sensor
Impacts
What was accomplished under these goals?
?A simple, lightweight, convolutional neural network ML model, 3 convolution layers deep, was found to be capable of distinguishing between minute differences in the shapes or the aspect ratios of RBCs. The model was trained on microscope images of cells immersed in an isotonic solution of a nematic liquid crystal (DSCG). We demonstrate that LCs biomaterial would 'sense' small differences in the mechanical properties of cells, as evident from the changed shape of the cell or the liquid crystalline color pattern around the cells (in the polarized images) and could be used to detect the presence/onset of diseases, or microbes in air or water in tandem with trained ML models. An underlying and important hypothesis - that biochemical changes in an organism could affect mechanical changes in its cells - makes for a rich and interesting endeavor for future researchers; similar to how the expression of HSP70 protein was shown to make the heat-stressed RBCs stiffer in this study. The 250K parameter model is easily trainable on GPU equipped personal computers using just a few hundred study-specific micrographs. The trained model is lightweight enough to fit in a sensor computer and fast enough to virtually instantly perform the classification on every new 'photograph'. In the context of the current rise of foundation vision models, whose fine-tuning and training are prohibitively expensive and possible only at very large research laboratories, our approach is geared towards empowering labs and individuals all round the world to construct and train their own ML models on general or niche tasks related to imaging and/or sensing. In experiments, it was found that crosslinking the chicken and human RBCs using glutaraldehyde in order to simulate a diseased cell was an adequate strategy for planning, building, training, and evaluating valid ML models ahead of collecting of actual training data. In our case, no model-tweaking was found to be necessary while going from the simulated to the real heat-stressed cells. Because biological data could often be available in less quantities and later in the study, we believe that our simulation example might come in handy for researchers looking to work on their ML models while waiting, or those looking to simply augment their data in an appropriate manner. The successful implementation of the CNN algorithm to rapidly identify HS also opens up new questions related to our line of investigation including: i) Are final static response (micrographs) enough to develop a ML classifier to discriminate (specify) different RBC strains? ii) How do the dynamic conditions (response of cells as a function of time under HS) contribute to developing efficient and specific classifiers? iii) Can static images at TN conditions predict the expression levels of HSP70 at HS conditions. Future work to answer these questions we will focus on collecting and maintaining a large and diverse image data base of strained and unstrained samples, and by developing dedicated deeper CNNs that are computationally efficient through use of ResNets. Additionally, CNN frameworks that process multi-channel image datasets, that include spectral channels, and dynamic images, beyond conventional 3 channel RGB images.
Publications
- Type:
Other Journal Articles
Status:
Published
Year Published:
2024
Citation:
A liquid crystal-based biomaterial platform for rapid sensing of heat stress using machine learning, P. Verma, E. Adeogun, E. S. Greene, S. Dridi, U. Nakarmi, K. Nayani*. Sensors and Diagnostics. 3 (11), 1843 (2024)
|
Progress 01/15/22 to 01/14/23
Outputs
Target Audience:In this project period our work was mainly disemminated via publications ( one published and one submitted). Therefore, the target audience was the populations engaging in scholary publications including scientists, farmers and agricultural researchers. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?Graudate student was trained in obtaining optical microgaphs and then in coding the CNN layers. How have the results been disseminated to communities of interest?We have published one manuscript in a scientific journal and currently have the second manuscript under review. What do you plan to do during the next reporting period to accomplish the goals?We aim to reach a 99% accuracy for the model in identifying heat stress inpreviously unseen optical data. We expect to work on tuning the hyper parameters to acheive this goal. Further, we aim to conduct new experiments on the specificity for the test and train a new model wherein there are several other stressors affecting the chicken
Impacts
What was accomplished under these goals?
This year, we were able to successfully accomplished the following sub-goals: Sample preparation and optical microscopy As-received RBCs were dispersed in a 154 mM NaCl isotonic solution; an isotonic solution ensured that cells maintained their natural elliptical shape. Typically, about 5 µL RBCs were mixed in about 60 µL isotonic NaCl solution to obtain the RBC dispersion. A 17.3% (w/v) DSCG aqueous solution was prepared by mixing DSCG in water in a vortex mixer for 4 hours. Previous works have showed that 17.3% (w/v) aqueous DSCG is isotonic with the interiors of an RBC [ref], which ensured that the RBC shape change was solely due to the mechanical interaction between the LC field and the RBC. Since RBCs naturally strain when put in a DSCG solution, strained cell samples were obtained by adding 2 μL of dispersed RBCs to 60 μL of DSCG solution and gently swirled. To prevent straining in DSCG and obtain unstrained cell samples, glutaraldehyde was used to crosslink and stiffen the RBCs [ref]. A stock solution of 5% v/v of glutaraldehyde in water was used; 5 µL RBCs were slowly pipetted into 0.2 µL of this stock to effect crosslinking. The final glutaraldehyde concentration in the cells was chosen to be around 0.2 % to make sure the individual cells were fixated and do not form aggregates. The solution was slowly mixed on a shaker for an hour to allow the glutaraldehyde to completely crosslink. About 2 µL of crosslinked RBC was then added to 60 µL of DSCG solution and gently swirled.HSP70 RBCs were collected from 21-day old broiler chickens that had been exposed to acute heat stress (35°C for 2 hours). Whole blood was collected into EDTA coated tubes and the RBCs were isolated from the whole blood by centrifugation and washing with PBS three times. For imaging, RBC samples were transferred (post swirling) to microscope slides. Micrographs were obtained using an Olympus BX41 optical microscope fitted with a 40x objective lens. Polarized and brightview images were captured in the presence and absence of a polarizer respectively. Building datasets Images from the microscope(s) were obtained in a variety of sizes, aspect ratios and formats. Images with differences in tint, brightness, contrast, lighting were included. Images with scratch marks on the microscope slides or of samples containing foreign objects (like dirt or lint) or containing things other than the RBCs were also included. This was done to increase the diversity in the dataset, keeping in mind the plausible diversity that the trained model may encounter during testing and after deployment. Collected images were at least 1800 px in height, either 3:2 or 4:3 in aspect ratio (width to heigh ratio), and saved in one of jpeg, bmp, tiff, or raw format. All images were RGB (containing information in red, green, and blue channels). Before building the dataset, all images were cropped and resized to the same size and converted to jpeg format. A square section from the center of the image was selected. This was done to (1) discard the sides which sometimes contained portion of the microscope slide or stage outside of the actual sample and (2) standardize the aspect ratio to 1:1 from 3:2 or 4:3. The resulting square image was scaled down in size to exactly 1000 px wide and 1000 px tall using bicubic resizing algorithm in PIL. Each image was assigned labels (such as species, chemicals used, and magnification etc.) that were stored in a tabular form within excel files. These labels were used to programmatically find images matching a certain criterion. For instance, polarized images of unstrained chicken RBCs were found by logical querying of these labels: polarized is "True", chemicals used contains "glutaraldehyde", species is "chicken", and cell type is "RBC". Ten distinct subsets of the dataset (called classes) were built using this process and are summarized in Table 1. Each class was shuffled and split into training (train), validation (val), and test (test) sets in 70:15:15 ratio, respectively. Table 1. Split of images between training, validation and test set in each dataset class or dataset name train val test brightview unstrained chicken 112 24 24 strained chicken 585 125 125 unstrained human 350 75 75 strained human 350 75 75 strained HS chicken 489 105 105 polarized unstrained chicken 798 171 171 strained chicken 1264 271 271 unstrained human 390 83 83 strained human 350 75 75 strained HS chicken 423 90 90 Building the CNN Convolutional neural networks (CNNs) were built using the TensorFlow library (that utilizes Keras API) in Python. Images from the datasets were imported using flow_from_directory function of Keras's preprocessing module, which yields batches of images indefinitely during training. The preprocessing module was also set to randomly flip the images horizontally or vertically during training, to add to the richness of the dataset as part of the data augmentation process. The batch size was set to 32 and the imported image size was set to 500 x 500 pixels. A simple binary classifier CNN, meaning it classified images into one of two input classes, was built. The model is schematically shown in Figure 2b. It consisted of three convolutional layers, each followed by a max-pooling layer. A ReLU activation function was used with each convolutional layer. Image tensors were flattened to a single dimension before being passed to the final dense layer consisting of a single neuron; this neuron yielded the probability that a particular image belonged to one of the two classes. This probability was yielded by a sigmoid activation function, that maps the input to a value between 0 and 1. The model's parameters are shown in Table 2. Model was compiled using the optimizer Adam and the loss function sparse_categorical_crossentropy. Performance of the model was measured using accuracy as the metric. Table 2. Summary of the CNN architecture obtained using Keras's summary method after defining the model. The type, shape of the output tensor, and the number of parameters (params) for each Keras layer are shown. layer name (type) output shape params conv2d_1 (Conv2D) (None, 498, 498, 16) 448 maxPooling-1 (MaxPooling2D) (None, 249, 249, 16) 0 conv2d_2 (Conv2D) (None, 247, 247, 32) 4640 maxPooling-2 (MaxPooling2D) (None, 123, 123, 32) 0 conv2d-3 (Conv2D) (None, 121, 121, 64) 18496 maxPooling-3 (MaxPooling2D) (None, 60, 60, 64) 0 flatten_1 (Flatten) (None, 230400) 0 dense_1 (Dense) (None, 1) 230401 Total params: 253,985 Trainable params: 253,985 Training the CNN In a given machine-learning experiment, a model could be trained to distinguish between two of any of the classes shown in Table 1. Specifically, models were trained to distinguish between unstrained and strained cells ('A and B', 'C and D', 'F and G', 'H and I') and between healthy and HS chicken cells ('B and E', 'G and J'). Experiments are depicted as first initials of class names (such as 'AB', 'CD' and so on) used in it, for convenience. Training was performed using the fit function on training image set, and accuracy was calculated and recorded during training for both training and validation sets. A receiver operating characteristic (ROC) curve was plotted for each experiment for a range of decision thresholds (101 linearly spaced threshold values from 0 to 1) for predictions made on validation sets. The best threshold was selected to be the one that was closest to the point (0, 1) on the ROC plot. The confusion matrix corresponding to the best threshold has been reported. A confusion matrix was also calculated for the predictions made on the test set using the best threshold (that was evaluated on the validation set) and has been reported.
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2022
Citation:
Effects of heat stress on cyto(chemo)kine and inflammasome gene expression and mechanical properties in isolated red and white blood cells from 4 commercial broiler lines and their ancestor jungle fowl. Elizabeth S. Greene, Elizabeth Adeogun, Sara K. Orlowski, Karthik Nayani, Sami Dridi. Poultry Science, Volume 101, Issue 6, June 2022, 101827
|