Development of a screening system of gene sets for estimating the time of early skeletal muscle injury based on second-generation sequencing technology.

The present study is aimed to address the challenge of wound age estimation in forensic science by identifying reliable genetic markers using low-cost and high-precision second-generation sequencing technology. A total of 54 Sprague-Dawley rats were randomly assigned to a control group or injury groups, with injury groups being further divided into time points (4 h, 8 h, 12 h, 16 h, 20 h, 24 h, 28 h, and 32 h after injury, n = 6) to establish rat skeletal muscle contusion models. Gene expression data were obtained using second-generation sequencing technology, and differential gene expression analysis, weighted gene co-expression network analysis (WGCNA) and time-dependent expression trend analysis were performed. A total of six sets of biomarkers were obtained: differentially expressed genes at adjacent time points (127 genes), co-expressed genes most associated with wound age (213 genes), hub genes exhibiting time-dependent expression (264 genes), and sets of transcription factors (TF) corresponding to the above sets of genes (74, 87, and 99 genes, respectively). Then, random forest (RF), support vector machine (SVM) and multilayer perceptron (MLP), were constructed for wound age estimation from the above gene sets. The results estimated by transcription factors were all superior to the corresponding hub genes, with the transcription factor group of WGCNA performed the best, with average accuracy rates of 96% for three models' internal testing, and 91.7% for the highest external validation. This study demonstrates the advantages of the indicator screening system based on second-generation sequencing technology and transcription factor level for wound age estimation.

Pathological Characteristics and Classification of Unstable Coronary Atherosclerotic Plaques.

Important forensic diagnostic indicators of sudden death in coronary atherosclerotic heart disease, such as acute or chronic myocardial ischemic changes, sometimes make it difficult to locate the ischemic site due to the short death process, the lack of tissue reaction time. In some cases, the deceased died of sudden death on the first-episode, resulting in difficulty for medical examiners to make an accurate diagnosis. However, clinical studies on coronary instability plaque revealed the key role of coronary spasm and thrombosis caused by their lesions in sudden coronary death process. This paper mainly summarizes the pathological characteristics of unstable coronary plaque based on clinical medical research, including plaque rupture, plaque erosion and calcified nodules, as well as the influencing factors leading to plaque instability, and briefly describes the research progress and technique of the atherosclerotic plaques, in order to improve the study on the mechanism of sudden coronary death and improve the accuracy of the forensic diagnosis of sudden coronary death by diagnosing different pathologic states of coronary atherosclerotic plaques.

Novel ratio-expressions of genes enables estimation of wound age in contused skeletal muscle.

Given that combination with multiple biomarkers may well raise the predictive value of wound age, it appears critically essential to identify new features under the limited cost. For this purpose, the present study explored whether the gene expression ratios provide unique time information as an additional indicator for wound age estimation not requiring the detection of new biomarkers and allowing full use of the available data. The expression levels of four wound-healing genes (Arid5a, Ier3, Stom, and Lcp1) were detected by real-time polymerase chain reaction, and a total of six expression ratios were calculated among these four genes. The results showed that the expression levels of four genes and six ratios of expression changed time-dependent during wound repair. The six expression ratios provided additional temporal information, distinct from the four genes analyzed separately by principal component analysis. The overall performance metrics for cross-validation and external validation of four typical prediction models were improved when six ratios of expression were added as additional input variables. Overall, expression ratios among genes provide temporal information and have excellent potential as predictive markers for wound age estimation. Combining the expression levels of genes with ratio-expression of genes may allow for more accurate estimates of the time of injury.

A macrophage plasma membrane-coated and DNA structured nanomedicine targets to alleviate rheumatoid arthritis via dual inhibition to TNF-α and NF-κB.

High heterogenicity of rheumatoid arthritis (RA) leads to poor response in many patients. Combined therapies that simultaneously inhibit multiple proinflammatory targets may improve anti-RA efficacy. However, which monotherapies to combine and how to achieve the combination are critical issues. Here, we design a macrophage plasma membrane-coated and DNA structured nanomedicine to achieve a dual inhibitory therapy to Tumor necrosis factor alpha (TNF-α) and NF-κB. An anti-NF-κB decoy oligodeoxynucleotides (dODN) is first conjugated to a DNA cage with precise numbers and locations (Cage-dODN). Meanwhile, an anti-TNF-α siRNA is anchored to extracted macrophage plasma membrane (siRNA@M). Subsequently, siRNA@M is used to encapsulate Cage-dODN to fabricate siRNA@M(Cage-dODN) (siMCO). The size and zeta potential of siMCO are 63.1 ± 15.7 nm and -20.7 ± 3.8 mV respectively. siMCO shows increased intracellular uptake by inflamed macrophages and enhanced accumulation in inflamed mouse paws. siMCO also reduces pro-inflammatory factors at genetic and protein levels, alleviates arthritic symptoms, and shows no influence to major blood components. These results show that siMCO is a potential targeted, efficient, and safe dual inhibitory therapy for the treatment of inflammatory arthritis. The macrophage plasma membrane can be utilized to improve the targeting, stability, and efficacy of DNA structured nanomedicines.

Combined metabolomics and tandem machine-learning models for wound age estimation: a novel analytical strategy.

Wound age estimation is one of the most challenging and indispensable issues for forensic pathologists. Although many methods based on physical findings and biochemical tests can be used to estimate wound age, an objective and reliable method for inferring the time interval after injury remains difficult. In the present study, endogenous metabolites of contused skeletal muscle were investigated to estimate the time interval after injury. Animal model of skeletal muscle injury was established using Sprague-Dawley rat, and the contused muscles were sampled at 4, 8, 12, 16, 20, 24, 28, 32, 36, 40, 44, and 48 h postcontusion (n = 9). Then, the samples were analysed using ultraperformance liquid chromatography coupled with high-resolution mass spectrometry. A total of 43 differential metabolites in contused muscle were determined by metabolomics method. They were applied to construct a two-level tandem prediction model for wound age estimation based on multilayer perceptron algorithm. As a result, all muscle samples were eventually divided into the following subgroups: 4, 8, 12, 16-20, 24-32, 36-40, and 44-48 h. The tandem model exhibited a robust performance and achieved a prediction accuracy of 92.6%, which was much higher than that of the single model. In summary, the multilayer perceptron-multilayer perceptron tandem machine-learning model based on metabolomics data can be used as a novel strategy for wound age estimation in future forensic casework. Key Points: The changes of metabolite profile were correlated with the time interval after injury in contused skeletal muscle.A panel of 43 endogenous metabolites screened by ultraperformance liquid chromatography coupled with high-resolution mass spectrometry could distinguish the wound ages.The multilayer perceptron (MLP) algorithm exhibited a robust performance in wound age estimation using metabolites.The combination of matabolomics and MLP-MLP tandem model could improve the accuracy of inferring the time interval after injury.

Postmortem Interval Estimation Using Protein Chip Technology Combined with Multivariate Analysis Methods.

OBJECTIVES: To estimate postmortem interval (PMI) by analyzing the protein changes in skeletal muscle tissues with the protein chip technology combined with multivariate analysis methods. METHODS: Rats were sacrificed for cervical dislocation and placed at 16 ℃. Water-soluble proteins in skeletal muscles were extracted at 10 time points (0 d, 1 d, 2 d, 3 d, 4 d, 5 d, 6 d, 7 d, 8 d and 9 d) after death. Protein expression profile data with relative molecular mass of 14 000-230 000 were obtained. Principal component analysis (PCA) and orthogonal partial least squares (OPLS) were used for data analysis. Fisher discriminant model and back propagation (BP) neural network model were constructed to classify and preliminarily estimate the PMI. In addition, the protein expression profiles data of human skeletal muscles at different time points after death were collected, and the relationship between them and PMI was analyzed by heat map and cluster analysis. RESULTS: The protein peak of rat skeletal muscle changed with PMI. The result of PCA combined with OPLS discriminant analysis showed statistical significance in groups with different time points (P<0.05) except 6 d, 7 d and 8 d after death. By Fisher discriminant analysis, the accuracy of internal cross-validation was 71.4% and the accuracy of external validation was 66.7%. The BP neural network model classification and preliminary estimation results showed the accuracy of internal cross-validation was 98.2%, and the accuracy of external validation was 95.8%. There was a significant difference in protein expression between 4 d and 25 h after death by the cluster analysis of the human skeletal muscle samples. CONCLUSIONS: The protein chip technology can quickly, accurately and repeatedly obtain water-soluble protein expression profiles in rats' and human skeletal muscles with the relative molecular mass of 14 000-230 000 at different time points postmortem. The establishment of multiple PMI estimation models based on multivariate analysis can provide a new idea and method for PMI estimation.

Exploration and Practice of the "One Combination, Two Highlights, Three Combinations, Four in One" Innovative Talents Training Mode in Forensic Medicine.

Talent is one of the basic and strategic supports for building a modern socialist country in all aspects. Since the 1980s, the establishment of forensic medicine major and the cultivation of innovative talents in forensic medicine have become hot topics in higher education in forensic medicine. Over the past 43 years, the forensic medicine team of Shanxi Medical University has adhered to the joint education of public security and colleges, and made collaborative innovation, forming a training mode of "One Combination, Two Highlights, Three Combinations, Four in One" for innovative talents in forensic medicine. It has carried out "5+3/X" integrated reform, and formed a relatively complete talent training innovation mode and management system in teaching, scientific research, identification, major, discipline, team, platform and cultural construction. It has made a historic contribution to China's higher forensic education, accumulated valuable experience for the construction of first-class major and first-class discipline of forensic medicine, and provided strong support for the construction of the national new forensic talent training system. The popularization of this training mode is conducive to the rapid and sustainable development of forensic science, and provides more excellent forensic talents for national building, regional social development and the discipline construction of forensic science.

Exploring postmortem succession of rat intestinal microbiome for PMI based on machine learning algorithms and potential use for humans.

The microbial communities may undergo a meaningful successional change during the progress of decay and decomposition that could aid in determining the post-mortem interval (PMI). However, there are still challenges to applying microbiome-based evidence in law enforcement practice. In this study, we attempted to investigate the principles governing microbial community succession during decomposition of rat and human corpse, and explore their potential use for PMI of human cadavers. A controlled experiment was conducted to characterize temporal changes in microbial communities associated with rat corpses as they decomposed for 30 days. Obvious differences of microbial community structures were observed among different stages of decomposition, especially between decomposition of 0-7d and 9-30d. Thus, a two-layer model for PMI prediction was developed based on the succession of bacteria by combining classification and regression models using machine learning algorithms. Our results achieved 90.48% accuracy for discriminating groups of PMI 0-7d and 9-30d, and yielded a mean absolute error of 0.580d within 7d decomposition and 3.165d within 9-30d decomposition. Furthermore, samples from human cadavers were collected to gain the common succession of microbial community between rats and humans. Based on the 44 shared genera of rats and humans, a two-layer model of PMI was rebuilt to be applied for PMI prediction of human cadavers. Accurate estimates indicated a reproducible succession of gut microbes across rats and humans. Together these results suggest that microbial succession was predictable and can be developed into a forensic tool for estimating PMI.

Engineering and characterization of carbohydrate-binding modules for imaging cellulose fibrils biosynthesis in plant protoplasts.

Carbohydrate binding modules (CBMs) are noncatalytic domains that assist tethered catalytic domains in substrate targeting. CBMs have therefore been used to visualize distinct polysaccharides present in the cell wall of plant cells and tissues. However, most previous studies provide a qualitative analysis of CBM-polysaccharide interactions, with limited characterization of engineered tandem CBM designs for recognizing polysaccharides like cellulose and limited application of CBM-based probes to visualize cellulose fibrils synthesis in model plant protoplasts with regenerating cell walls. Here, we examine the dynamic interactions of engineered type-A CBMs from families 3a and 64 with crystalline cellulose-I and phosphoric acid swollen cellulose. We generated tandem CBM designs to determine various characteristic properties including binding reversibility toward cellulose-I using equilibrium binding assays. To compute the adsorption (nkon ) and desorption (koff ) rate constants of single versus tandem CBM designs toward nanocrystalline cellulose, we employed dynamic kinetic binding assays using quartz crystal microbalance with dissipation. Our results indicate that tandem CBM3a exhibited the highest adsorption rate to cellulose and displayed reversible binding to both crystalline/amorphous cellulose, unlike other CBM designs, making tandem CBM3a better suited for live plant cell wall biosynthesis imaging applications. We used several engineered CBMs to visualize Arabidopsis thaliana protoplasts with regenerated cell walls using confocal laser scanning microscopy and wide-field fluorescence microscopy. Lastly, we also demonstrated how CBMs as probe reagents can enable in situ visualization of cellulose fibrils during cell wall regeneration in Arabidopsis protoplasts.

Forensic identification of sudden cardiac death: a new approach combining metabolomics and machine learning.

The determination of sudden cardiac death (SCD) is one of the difficult tasks in the forensic practice, especially in the absence of specific morphological changes in the autopsies and histological investigations. In this study, we combined the metabolic characteristics from corpse specimens of cardiac blood and cardiac muscle to predict SCD. Firstly, ultra-high performance liquid chromatography coupled with high-resolution mass spectrometry (UPLC-HRMS)-based untargeted metabolomics was applied to obtain the metabolomic profiles of the specimens, and 18 and 16 differential metabolites were identified in the cardiac blood and cardiac muscle from the corpses of those who died of SCD, respectively. Several possible metabolic pathways were proposed to explain these metabolic alterations, including the metabolism of energy, amino acids, and lipids. Then, we validated the capability of these combinations of differential metabolites to distinguish between SCD and non-SCD through multiple machine learning algorithms. The results showed that stacking model integrated differential metabolites featured from the specimens showed the best performance with 92.31% accuracy, 93.08% precision, 92.31% recall, 91.96% F1 score, and 0.92 AUC. Our results revealed that the SCD metabolic signature identified by metabolomics and ensemble learning in cardiac blood and cardiac muscle has potential in SCD post-mortem diagnosis and metabolic mechanism investigations.

Novel Prediction Method Applied to Wound Age Estimation: Developing a Stacking Ensemble Model to Improve Predictive Performance Based on Multi-mRNA.

(1) Background: Accurate diagnosis of wound age is crucial for investigating violent cases in forensic practice. However, effective biomarkers and forecast methods are lacking. (2) Methods: Samples were collected from rats divided randomly into control and contusion groups at 0, 4, 8, 12, 16, 20, and 24 h post-injury. The characteristics of concern were nine mRNA expression levels. Internal validation data were used to train different machine learning algorithms, namely random forest (RF), support vector machine (SVM), multilayer perceptron (MLP), gradient boosting (GB), and stochastic gradient descent (SGD), to predict wound age. These models were considered the base learners, which were then applied to developing 26 stacking ensemble models combining two, three, four, or five base learners. The best-performing stacking model and base learner were evaluated through external validation data. (3) Results: The best results were obtained using a stacking model of RF + SVM + MLP (accuracy = 92.85%, area under the receiver operating characteristic curve (AUROC) = 0.93, root-mean-square-error (RMSE) = 1.06 h). The wound age prediction performance of the stacking models was also confirmed for another independent dataset. (4) Conclusions: We illustrate that machine learning techniques, especially ensemble algorithms, have a high potential to be used to predict wound age. According to the results, the strategy can be applied to other types of forensic forecasts.

GPR65 as a potential immune checkpoint regulates the immune microenvironment according to pan-cancer analysis.

It has been reported that inhibition of GPR65 may be effective for the treatment of certain cancers. Nevertheless, the role of GPR65 in various cancers remains unknown. We conducted an exhaustive pan-cancer analysis of GPR65 using multiple databases, including TCGA, GTEx, BioGPS, HPA, cBioPortal, and GeneCards. GPR65 was found to be differentially expressed in various cancers and linked to tumor mutational burden (TMB), microsatellite instability (MSI), and Ploidy, playing a key function in the tumor microenvironment (TME). It is closely linked to the development of Th17 cells as well as Th1 and Th2 cells in certain cancers. Our findings indicate that the expression of GPR65 is highly linked with clinical prognosis, mutations, and immune cell infiltration. It was revealed as an indicator of patient prognosis as well as a possible immunomodulatory role. As a possible new immunological checkpoint, GPR65 could be a target for tumor immunotherapy.

A Wireless Headstage System Based on Neural-Recording Chip Featuring 315 nW Kickback-Reduction SAR ADC.

Wireless neural-recording instruments eliminate the bulky cables in multi-channel signal transmission, while the system size should be reduced to mitigate the impact on freely-moving animals. As the battery usually dominates the system size, the neural-recording chip should be low power to minimize the battery in long-termly monitoring. In general, a neural-recording chip consists of an analog front end (AFE) and an 8 bit -10 bit analog-to-digital converter (ADC), while it's challenging to design an ADC with an 8 -10 effective number of bits (ENOB) and sub- µ W power consumption due to the kickback noise. In this work, we propose a kickback-reduction technique for a successive-approximation-register (SAR) ADC based on neural-recording chip. Fabricated in 65 nm CMOS process, the proposed technique reduce the ADC power to 315 nW, resulting in an 8-channel neural-recording chip with 249 µW in total. Measured results show that the chip achieves an ADC ENOB of 9.73 bits, as well as an AFE gain of 43.3 dB and input-referred noise (IRN) of 9.68 µVrms in a bandwidth of 0.9 Hz -7.2 kHz. Combined with a BLE chip and a PCB antenna, the chip is implemented into a 2.6 g wireless headstage system (w/o battery), and an in-vivo demonstration is conducted on a male Sprague-Dawley rat with Parkinson's disease. The headstage system transfers the in-vivo neural signals to a commodity smartphone through BLE, and the miniature size induces little impact on freely-moving activities.

A novel method for determining postmortem interval based on the metabolomics of multiple organs combined with ensemble learning techniques.

Determining postmortem interval (PMI) is one of the most challenging and essential endeavors in forensic science. Developments in PMI estimation can take advantage of machine learning techniques. Currently, applying an algorithm to obtain information on multiple organs and conducting joint analysis to accurately estimate PMI are still in the early stages. This study aimed to establish a multi-organ stacking model that estimates PMI by analyzing differential compounds of four organs in rats. In a total of 140 rats, skeletal muscle, liver, lung, and kidney tissue samples were collected at each time point after death. Ultra-performance liquid chromatography coupled with high-resolution mass spectrometry was used to determine the compound profiles of the samples. The original data were preprocessed using multivariate statistical analysis to determine discriminant compounds. In addition, three interrelated and increasingly complex patterns (single organ optimal model, single organ stacking model, multi-organ stacking model) were established to estimate PMI. The accuracy and generalized area under the receiver operating characteristic curve of the multi-organ stacking model were the highest at 93% and 0.96, respectively. Only 1 of the 14 external validation samples was misclassified by the multi-organ stacking model. The results demonstrate that the application of the multi-organ combination to the stacking algorithm is a potential forensic tool for the accurate estimation of PMI.

Combined metabolomics and machine learning algorithms to explore metabolic biomarkers for diagnosis of acute myocardial ischemia.

Acute myocardial ischemia (AMI) remains the leading cause of death worldwide, and the post-mortem diagnosis of AMI represents a current challenge for both clinical and forensic pathologists. In the present study, the untargeted metabolomics based on ultra-performance liquid chromatography combined with high-resolution mass spectrometry was applied to analyze serum metabolic signatures from AMI in a rat model (n = 10 per group). A total of 28 endogenous metabolites in serum were significantly altered in AMI group relative to control and sham groups. A set of machine learning algorithms, namely gradient tree boosting (GTB), support vector machine (SVM), random forest (RF), logistic regression (LR), and multilayer perceptron (MLP) models, was used to screen the more valuable metabolites from 28 metabolites to optimize the biomarker panel. The results showed that classification accuracy and performance of MLP model were better than other algorithms when the metabolites consisting of L-threonic acid, N-acetyl-L-cysteine, CMPF, glycocholic acid, L-tyrosine, cholic acid, and glycoursodeoxycholic acid. Finally, 17 blood samples from autopsy cases were applied to validate the classification model's value in human samples. The MLP model constructed based on rat dataset achieved accuracy of 88.23%, and ROC of 0.89 for predicting AMI type II in autopsy cases of sudden cardiac death. The results demonstrated that MLP model based on 7 molecular biomarkers had a good diagnostic performance for both AMI rats and autopsy-based blood samples. Thus, the combination of metabolomics and machine learning algorithms provides a novel strategy for AMI diagnosis.

Application of Linear Regression Model of Gpnmb Gene in Rat Injury Time Estimation.

OBJECTIVES: To investigate the effects of injury time, postmortem interval (PMI) and postmortem storage temperature on mRNA expression of glycoprotein non-metastatic melanoma protein B (Gpnmb), and to establish a linear regression model between Gpnmb mRNA expression and injury time, to provide aimed at providing potential indexes for injury time estimation. METHODS: Test group SD rats were anesthetized and subjected to blunt contusion and randomly divided into 0 h, 4 h, 8 h, 12 h, 16 h, 20 h and 24 h groups after injury, with 18 rats in each group. After cervical dislocation, 6 rats in each group were collected and stored at 0 ℃, 16 ℃ and 26 ℃, respectively. The muscle tissue samples of quadriceps femoris injury were collected at 0 h, 12 h and 24 h postmortem at the same temperature. The grouping method and treatment method of the rats in the validation group were the same as above. The expression of Gpnmb mRNA in rat skeletal muscle was detected by RT-qPCR. The Pearson correlation coefficient was used to evaluate the correlation between Gpnmb mRNA expression and injury time, PMI, and postmortem storage temperature. SPSS 25.0 software was used to construct a linear regression model, and the validation group data was used for the back-substitution test. RESULTS: The expression of Gpnmb mRNA continued to increase with the prolongation of injury time, and the expression level was highly correlated with injury time (P<0.05), but had little correlation with PMI and postmortem storage temperature (P>0.05). The linear regression equation between injury time (y) and Gpnmb mRNA relative expression (x) was y=0.611 x+4.489. The back-substitution test proved that the prediction of the model was accurate. CONCLUSIONS: The expression of Gpnmb mRNA is almost not affected by the PMI and postmortem storage temperature, but is mainly related to the time of injury. Therefore, a linear regression model can be established to infer the time of injury.

Analysis of sensitivity and specificity: precise recognition of neutrophils during regeneration of contused skeletal muscle in rats.

In this report, we applied the TissueFAXS 200 digital pathological analysis system to rapidly and accurately identify neutrophils during regeneration of contused skeletal muscle, and to provide information for follow-up studies on neutrophils to estimate wound age. Rat injury model was established, and skeletal muscle samples were obtained from the control group and contusion groups at 1, 1.5, 2, 3, 4, and 6 h, as well as at 1, 3, 5, and 15 d post-injury (n = 5 per group). The expression of nuclei and neutrophils was detected by hematoxylin and eosin (HE) staining and immunohistochemical (IHC) staining. A total of 20 injury site areas of 0.25 mm2 (0.5 mm × 0.5 mm) were then randomly selected at all time points. A TissueFAXS 200 digital pathological analysis system was used to identify the positive and negative numbers. Knowledge of five professional medical workers were considered the gold standard to measure the false positive rate (FPR), false negative rate (FNR), sensitivity, specificity, and area under the curve (AUC) of receiver operating characteristic (ROC) curves. As a result, with a staining area of neutrophils from 8 µm2 to 15 µm2, the FPR was 4.28%-12.14%, the FNR was 12.42%-64.08%, the sensitivity was 35.92%-87.58%, the specificity was 87.86%-95.72%, the Youden index was 0.316-0.754, the accuracy was 82.80%-88.30%, and the AUC was 0.771-0.826. The AUC was largest when the cut-off value of the staining area was 12 µm2. Our results show that this software-based method is more accurate than the human eye in evaluating neutrophil infiltration. Based on the sensitivity and specificity, neutrophils can be accurately identified during regeneration of contused skeletal muscle. The TissueFAXS 200 digital pathological analysis system can also be used to optimize conditions for different cell types under various injury conditions to determine the optimal cut-off value of the staining area and provide optimal conditions for further study. Furthermore, it will provide evidence for forensic pathology cases.

Comparison of Medical Dispute Resolution Mechanisms in China and Abroad.

Medical disputes are one of the common problems concerned by the whole world. All countries and regions have established their own medical dispute resolution mechanisms, in accordance with their own national conditions. Medical dispute identification opinions, as one of the important bases for identifying the responsibilities of both doctors and patients, play a pivotal role in the process of dispute settlement. A reasonable medical dispute resolution mechanism and standardized medical dispute identification model can help resolve disputes flexibly and reduce the conflict between doctors and patients. This paper briefly compares the medical dispute resolution mechanism and identification mode of China and several other representative countries (the United States, Britain, France, Germany, Italy, Japan, etc.), and discusses their respective characteristics and shortcomings, to bring some enlightenment to the medical dispute resolution and identification in our country.

Research Progress of Automatic Diatom Test by Artificial Intelligence.

Diatom test is the main laboratory test method in the diagnosis of drowning in forensic medicine. It plays an important role in differentiating the antemortem drowning from the postmortem drowning and inferring drowning site. Artificial intelligence (AI) automatic diatom test is a technological innovation in forensic drowning diagnosis which is based on morphological characteristics of diatom, the application of AI algorithm to automatic identification and classification of diatom in tissues and organs. This paper discusses the morphological diatom test methods and reviews the research progress of automatic diatom recognition and classification involving AI algorithms. AI deep learning algorithm can assist diatom testing to obtain objective, accurate, and efficient qualitative and quantitative analysis results, which is expected to become a new direction of diatom testing research in the drowning of forensic medicine in the future.

Comparison among Four Deep Learning Image Classification Algorithms in AI-based Diatom Test.

OBJECTIVES: To select four algorithms with relatively balanced complexity and accuracy among deep learning image classification algorithms for automatic diatom recognition, and to explore the most suitable classification algorithm for diatom recognition to provide data reference for automatic diatom testing research in forensic medicine. METHODS: The "diatom" and "background" small sample size data set (20 000 images) of digestive fluid smear of corpse lung tissue in water were built to train, validate and test four convolutional neural network (CNN) models, including VGG16, ResNet50, InceptionV3 and Inception-ResNet-V2. The receiver operating characteristic curve (ROC) of subjects and confusion matrixes were drawn, recall rate, precision rate, specificity, accuracy rate and F1 score were calculated, and the performance of each model was systematically evaluated. RESULTS: The InceptionV3 model achieved much better results than the other three models with a balanced recall rate of 89.80%, a precision rate of 92.58%. The VGG16 and Inception-ResNet-V2 had similar diatom recognition performance. Although the performance of diatom recall and precision detection could not be balanced, the recognition ability was acceptable. ResNet50 had the lowest diatom recognition performance, with a recall rate of 55.35%. In terms of feature extraction, the four models all extracted the features of diatom and background and mainly focused on diatom region as the main identification basis. CONCLUSIONS: Including the Inception-dependent model, which has stronger directivity and targeting in feature extraction of diatom. The InceptionV3 achieved the best performance on diatom identification and feature extraction compared to the other three models. The InceptionV3 is more suitable for daily forensic diatom examination.