Efficiently improving the Wi-Fi-based human activity recognition, using auditory features, autoencoders, and fine-tuning.

Human activity recognition (HAR) based on Wi-Fi signals has attracted significant attention due to its convenience and the availability of infrastructures and sensors. Channel State Information (CSI) measures how Wi-Fi signals propagate through the environment. However, many scenarios and applications have insufficient training data due to constraints such as cost, time, or resources. This poses a challenge for achieving high accuracy levels with machine learning techniques. In this study, multiple deep learning models for HAR were employed to achieve acceptable accuracy levels with much less training data than other methods. A pretrained encoder trained from a Multi-Input Multi-Output Autoencoder (MIMO AE) on Mel Frequency Cepstral Coefficients (MFCC) from a small subset of data samples was used for feature extraction. Then, fine-tuning was applied by adding the encoder as a fixed layer in the classifier, which was trained on a small fraction of the remaining data. The evaluation results (K-fold cross-validation and K = 5) showed that using only 30% of the training and validation data (equivalent to 24% of the total data), the accuracy was improved by 17.7% compared to the case where the encoder was not used (with an accuracy of 79.3% for the designed classifier, and an accuracy of 90.3% for the classifier with the fixed encoder). While by considering more calculational cost, achieving higher accuracy using the pretrained encoder as a trainable layer is possible (up to 2.4% improvement), this small gap demonstrated the effectiveness and efficiency of the proposed method for HAR using Wi-Fi signals.

CSI-Based Human Activity Recognition Using Multi-Input Multi-Output Autoencoder and Fine-Tuning.

Wi-Fi-based human activity recognition (HAR) has gained considerable attention recently due to its ease of use and the availability of its infrastructures and sensors. Channel state information (CSI) captures how Wi-Fi signals are transmitted through the environment. Using channel state information of the received signals transmitted from Wi-Fi access points, human activity can be recognized with more accuracy compared with the received signal strength indicator (RSSI). However, in many scenarios and applications, there is a serious limit in the volume of training data because of cost, time, or resource constraints. In this study, multiple deep learning models have been trained for HAR to achieve an acceptable accuracy level while using less training data compared to other machine learning techniques. To do so, a pretrained encoder which is trained using only a limited number of data samples, is utilized for feature extraction. Then, by using fine-tuning, this encoder is utilized in the classifier, which is trained by a fraction of the rest of the data, and the training is continued alongside the rest of the classifier's layers. Simulation results show that by using only 50% of the training data, there is a 20% improvement compared with the case where the encoder is not used. We also showed that by using an untrainable encoder, an accuracy improvement of 11% using 50% of the training data is achievable with a lower complexity level.

Comparison of PCR-HRM, colorimetric LAMP and culture based diagnostic assays in the detection of endometritis caused by Streptococcus equi subsp. zooepidemicus in mares.

Streptococcus equi subsp. zooepidemicus (S. zooepidemicus) is one of the causative agents of equine endometritis. In this study, a panel of different bacterial species, and colonies derived from bacteriological cultures of 38 clinical samples, were subjected to Loop-Mediated Isothermal Amplification (LAMP) assay and PCR, followed by high-resolution melt (HRM) curve analysis. All clinical samples were genotyped into three distinct groups based on HRM curve analysis. Differences in melting curve profiles were a reflection of DNA variation in sorD gene which was confirmed by DNA sequencing. A mathematical model based on Genetic Confidence Percentage (GCP) was used in HRM curve analysis and a cut-off point value was established which differentiated S. zooepidemicus isolates without requiring visual interpretation of curve profiles. The accuracy of PCR-HRM and bacterial culture in detection of S. zooepidemicus were identical with 100% sensitivity and specificity, while LAMP assay had similar specificity but a lower sensitivity (89.5%). PCR-HRM and LAMP assay provided an effective detection method with a turn-around time of six hours for PCR-HRM and 120 min for LAMP assay, compared to a minimum three days that was required when routine bacteriological culture method was used. In summary, results indicate that LAMP had the quickest turnaround, and HRM curve analysis could potentially be used for genotyping without DNA sequencing. Any mare suspected of endometritis will benefit from developed rapid diagnostic tests for detection of S. zooepidemicus and proper treatment prior to being bred and will mitigate unnecessary treatment and antibiotic resistance.

Comparison of colourimetric loop-mediated isothermal amplification (LAMP), PCR and high-resolution melt curve analysis and culture-based diagnostic assays in the detection of three salmonella serotypes in poultry.

The accuracies of two molecular tests, PCR and loop-mediated isothermal amplification (LAMP) assay were compared with bacterial culture in detection of salmonella in poultry clinical samples. The icIR family transcriptional regulator gene was targeted and, out of 56 clinical specimens, 20 poultry field isolates were found positive for salmonella. Along with human isolates, reference strains of three different serovars, Salmonella Enteritidis (S. Enteritidis), S. Typhimurium and S. Infantis, were also tested. Eight different but genetically closely related bacterial genera (Klebsiella, Pseudomonas, Enterobacter, Campylobacter, Staphylococcus, Streptococcus, Escherichia and Pasteurella) were also used to evaluate the specificity of assay. The LAMP assay showed 80.8% sensitivity (95% CI, 0.66-0.95) and 100% specificity (95% CI, 0.71-1.00) when compared with microbiological culture and PCR, both with 100% sensitivity (95% CI, 0.87-1.00) and 100% specificity (95% CI, 0.71-1.00). High-resolution melt (HRM) curve analysis following PCR was able to differentiate between salmonella isolates based on their melting points, and all specimens were genotyped in three distinct HRM curve profiles. Each normalized melt curve profile represented one salmonella serotype and differences between the three melt profiles were correlated with nucleotide variations in the target gene sequences which demonstrated high discriminatory power of this technique. The colourimetric LAMP assay provided an alternative detection method capable of being used in the field, and showed analytical sensitivity for detection of 1 pg of salmonella DNA per reaction. The advantages and disadvantages of each test in detection of salmonella are discussed.

A CSI-Based Human Activity Recognition Using Deep Learning.

The Internet of Things (IoT) has become quite popular due to advancements in Information and Communications technologies and has revolutionized the entire research area in Human Activity Recognition (HAR). For the HAR task, vision-based and sensor-based methods can present better data but at the cost of users' inconvenience and social constraints such as privacy issues. Due to the ubiquity of WiFi devices, the use of WiFi in intelligent daily activity monitoring for elderly persons has gained popularity in modern healthcare applications. Channel State Information (CSI) as one of the characteristics of WiFi signals, can be utilized to recognize different human activities. We have employed a Raspberry Pi 4 to collect CSI data for seven different human daily activities, and converted CSI data to images and then used these images as inputs of a 2D Convolutional Neural Network (CNN) classifier. Our experiments have shown that the proposed CSI-based HAR outperforms other competitor methods including 1D-CNN, Long Short-Term Memory (LSTM), and Bi-directional LSTM, and achieves an accuracy of around 95% for seven activities.

Evaluation of Two Multiplex PCR-High-Resolution Melt Curve Analysis Methods for Differentiation of Campylobacter jejuni and Campylobacter coli Intraspecies.

Campylobacter infection is a common cause of bacterial gastroenteritis in humans and remains a significant global public health issue. The capability of two multiplex PCR (mPCR)-high-resolution melt (HRM) curve analysis methods (i.e., mPCR1-HRM and mPCR2-HRM) to detect and differentiate 24 poultry isolates and three reference strains of Campylobacter jejuni and Campylobacter coli was investigated. Campylobacter jejuni and C. coli were successfully differentiated in both assays, but the differentiation power of mPCR2-HRM targeting the cadF gene was found superior to that of mPCR1-HRM targeting the gpsA gene or a hypothetical protein gene. However, higher intraspecies variation within C. coli and C. jejuni isolates was detected in mPCR1-HRM when compared with mPCR2-HRM. Both assays were rapid and required minimum interpretation skills for discrimination between and within Campylobacter species when using HRM curve analysis software.

Maximum Likelihood Estimation for Multiple Camera Target Tracking on Grassmann Tangent Subspace.

In this paper, we introduce a likelihood model for tracking the location of object in multiple view systems. Our proposed model transforms conventional nonlinear Euclidean estimation model to an estimation model based on the manifold tangent subspace. In this paper, we show that by decomposition of input noise into two parts and description of model by exponential map, real observations in the Euclidean geometry can be transformed to the manifold tangent subspace. Moreover, by obtained tangent subspace likelihood function, we propose two iterative and noniterative maximum likelihood estimation approaches which numerical results show their good performance.

The complete mitochondrial genome of Linguatula serrata (tongue worm) isolated from a dog and phylogenetic analysis.

The complete mitogenome of Linguatula serrata isolated from nasal cavity of a dog was characterized for the first time. The total size of the circular mitogenome was 15,328 bp consisting of 37 genes including 13 protein coding genes, 22 tRNA genes, two rRNA genes and two control regions. Phylogenetic tree was constructed based on 17 closely related species and their genetic relationship with Linguatula serrata was analysed.

Evaluation of PCR and high-resolution melt curve analysis for differentiation of Salmonella isolates.

Consumption of poultry products contaminated with Salmonella is one of the major causes of foodborne diseases worldwide and therefore detection and differentiation of Salmonella spp. in poultry is important. In this study, oligonucleotide primers were designed from hemD gene and a PCR followed by high-resolution melt (HRM) curve analysis was developed for rapid differentiation of Salmonella isolates. Amplicons of 228 bp were generated from 16 different Salmonella reference strains and from 65 clinical field isolates mainly from poultry farms. HRM curve analysis of the amplicons differentiated Salmonella isolates and analysis of the nucleotide sequence of the amplicons from selected isolates revealed that each melting curve profile was related to a unique DNA sequence. The relationship between reference strains and tested specimens was also evaluated using a mathematical model without visual interpretation of HRM curves. In addition, the potential of the PCR-HRM curve analysis was evaluated for genotyping of additional Salmonella isolates from different avian species. The findings indicate that PCR followed by HRM curve analysis provides a rapid and robust technique for genotyping of Salmonella isolates to determine the serovar/serotype.

A comparison of PCR assays for beak and feather disease virus and high resolution melt (HRM) curve analysis of replicase associated protein and capsid genes.

Beak and feather disease virus (BFDV) threatens a wide range of endangered psittacine birds worldwide. In this study, we assessed a novel PCR assay and genetic screening method using high-resolution melt (HRM) curve analysis for BFDV targeting the capsid (Cap) gene (HRM-Cap) alongside conventional PCR detection as well as a PCR method that targets a much smaller fragment of the virus genome in the replicase initiator protein (Rep) gene (HRM-Rep). Limits of detection, sensitivity, specificity and discriminatory power for differentiating BFDV sequences were compared. HRM-Cap had a high positive predictive value and could readily differentiate between a reference genotype and 17 other diverse BFDV genomes with more discriminatory power (genotype confidence percentage) than HRM-Rep. Melt curve profiles generated by HRM-Cap correlated with unique DNA sequence profiles for each individual test genome. The limit of detection of HRM-Cap was lower (2×10-5ng/reaction or 48 viral copies) than that for both HRM-Rep and conventional BFDV PCR which had similar sensitivity (2×10-6ng or 13 viral copies/reaction). However, when used in a diagnostic setting with 348 clinical samples there was strong agreement between HRM-Cap and conventional PCR (kappa=0.87, P<0.01, 98% specificity) and HRM-Cap demonstrated higher specificity (99.9%) than HRM-Rep (80.3%).

Phylogenetic relationships among Linguatula serrata isolates from Iran based on 18S rRNA and mitochondrial cox1 gene sequences.

The phylogenetic relationships among seven Linguatula serrata (L. serrata) isolates collected from cattle, goats, sheep, dogs and camels in different geographical locations of Iran were investigated using partial 18S ribosomal RNA (rRNA) and partial mitochondrial cytochrome c oxidase subunit 1 (cox1) gene sequences. The nucleotide sequences were analysed in order to determine the phylogenetic relationships between the isolates. Higher sequence diversity and intraspecies variation was observed in the cox1 gene compared to 18S rRNA sequences. Phylogenetic analysis of the cox1 gene placed all L. serrata isolates in a sister clade to L. arctica. The Mantel regression analysis revealed no association between genetic variations and host species or geographical location, perhaps due to the small sample size. However, genetic variations between L. serrata isolates in Iran and those isolated in other parts of the world may exist and could reveal possible evolutionary relationships.

Evolutionary characterization of hemagglutinin gene of H9N2 influenza viruses isolated from Asia.

The full length hemagglutinin (HA) genes of 287 H9N2 AI strains isolated from chickens in Asia during the period 1994-2009 were genetically analyzed. Phylogenetic analysis showed that G1-like viruses circulated in the Middle East and Indian sub-continent countries, whereas other sublineages existed in Far East countries. It also revealed G1-like viruses with an average 96.7% identity clustered into two subgroups largely based on their time of isolation. The Ka/Ks ratio was calculated 0.34 for subgroup 1 and 0.57 for subgroup 2 indicates purifying/stabilizing selection, but despite this there is evidence of localized positive selection when comparing the subgroups 1 and 2 protein sequences. Five sites in HA H9N2 viruses had a posterior probability >0.5 using the Bayesian method, indicating these sites were under positive selection. These sites were found to be associated with the globular head region of HA. To identify sites under positive selection; amino acid substitution classified depends on their radicalism and neutrality. The results indicate that, although most positions in HAs were under purifying selection and can be eliminated, a few positions located in the antigenic regions and receptor binding sites were subject to positive selection.

Specific subtyping of influenza A virus using a recombinant hemagglutinin protein expressed in baculovirus.

Influenza A viruses are subtyped according to antigen characterization of hemagglutinin (HA) and neuraminidase surface glycoproteins. The hemagglutination inhibition (HI) assay using reference antiserum is currently applied to serologic screening of subtype-specific antibodies in sera. The reference antiserum is made by injecting chickens with live or inactivated whole virus preparations. Nonspecific inhibitors of antisera prepared by the conventional method may affect the specificity of HI assay. In this study, highly pure recombinant proteins generated using baculovirus expression vector system based on full-length of HA (HAF) and antigenic region of HA(1) genes of H9 subtype, and also inactivated whole virus were used to immunization of chickens. Measurable antibody titers were present for treated birds after 3 weeks and generally increased after each boost. The performance of the prepared antisera was evaluated by testing a panel of known standard strains of influenza virus representing five HA subtypes. Relative to the conventional method using whole virus immunization and recombinant HAF protein, the antiserum prepared by recombinant HA(1) had a specificity of 100% for all tested subtypes. The antiserum prepared by expression of HA1 protein in baculovirus has the potential for rapid and specific HA subtyping of influenza viruses without producing antibodies specific to other viral proteins.

Development of rHA1-ELISA for specific and sensitive detection of H9 subtype influenza virus.

A recombinant antigen-based single serum dilution ELISA was developed for simultaneous detection and subtyping of influenza viruses. Recombinant baculovirus encoding the hemagglutinin (HA(1) subunit) of H9N2 virus was generated. To evaluate the rHA1-ELISA, microplates were coated with purified HA1 protein and tested with reference control sera. Subsequently, 92 field sera collected from chickens suspected to be infected with H9N2 AIV were employed to test the efficacy of the rHA1-ELISA. The sera were tested simultaneously by HI and a commercial AIV ELISA kit. The rHA1-ELISA appeared to be highly specific and sensitive for direct detection of H9N2 antibodies in serum samples.

Ethnic and geographic differentiation of Helicobacter pylori within Iran.

The bacterium Helicobacter pylori colonizes the human stomach, with individual infections persisting for decades. The spread of the bacterium has been shown to reflect both ancient and recent human migrations. We have sequenced housekeeping genes from H. pylori isolated from 147 Iranians with well-characterized geographical and ethnic origins sampled throughout Iran and compared them with sequences from strains from other locations. H. pylori from Iran are similar to others isolated from Western Eurasia and can be placed in the previously described HpEurope population. Despite the location of Iran at the crossroads of Eurasia, we found no evidence that the region been a major source of ancestry for strains across the continent. On a smaller scale, we found genetic affinities between the H. pylori isolated from particular Iranian populations and strains from Turks, Uzbeks, Palestinians and Israelis, reflecting documented historical contacts over the past two thousand years.

A clinic-based screening of mutations in exons 31, 34, 35, 41, and 48 of LRRK2 in Iranian Parkinson's disease patients.

We present results of mutation screening of exons 31, 34, 35, 41, and 48 of LRRK2 in 205 Iranian Parkinson's disease patients. Sixteen percent of the cases were familial. Although age was not a factor in patient recruitment, the Iranian cohort was relatively young (average age at onset of disease: 48.9 years). A notably high male to female ratio (2.96:1) and earlier age at onset (by 2.9 years) in men were observed. A known disease-associated variation, c.C4321T causing R1441C, and IVS31 + 3A > G, a variation that may be associated, were observed. Therefore, disregarding IVS31 + 3A > G, disease status in at least 0.5% of our young cohort and in 3.5% of the familial cases was associated with a mutation in the five exons of LRRK2 screened. Interestingly, the variation causing p.G2019S was not observed.

PCR based RFLP genotyping of bovine lymphocyte antigen DRB3.2 in Iranian Holstein population.

Major Histocompatibility Complex (MHC) class II locus DRB3 was investigated by PCR based restriction fragment length polymorphism (PCR-RFLP) assay. A total of 262 Holstein cows participating in the national recording system were sampled from 10 herds. A two-step polymerase chain reaction was carried out in order to amplify a 284 base-pair fragment of exon 2 of the target gene. Second PCR products were treated with three restriction endonucleas enzymes RsaI, BstYI and HaeIII. Digested fragments were analyzed by polyacrylamid gel electrophoresis. Twenty-eight BoLA-DRB3 alleles were identified. Identified alleles are: BoLA-DRB3.2 *3, *6, *7, *8, *9, *10, *11, *12, *13, *14, *15, *16,20, *21, *22, *23, *24, *25, *26, *27, *28, *32, *36, *37, *40, *51, *iaa and *ibb. The BoLA-DRB3.2*40 allele that was observed in this study has not been reported previously. The calculated frequencies were as follows: 2.29, 1.34, 0.19, 14.5, 0.38, 3.05, 12.21, 1.34, 2.29, 1.34, 2.48, 9.16, 0.95, 0.77, 6.68, 9.16, 17.94, 1.15, 0.57, 1.15, 0.95, 0.57, 0.38, 1.91, 0.38, 5.73, 0.19 and 0.95% respectively. The six most frequently observed alleles (BoLA-DRB3.2 *8, *11, *16, *22, *23 and *24) accounted for 69.65% of the alleles in these 10 herds. The results of this study confirm the allelic distribution of six most frequent alleles in Holstein population's worldwide.