An optimised YOLOv4 deep learning model for efficient malarial cell detection in thin blood smear images.

BACKGROUND: Malaria is a serious public health concern worldwide. Early and accurate diagnosis is essential for controlling the disease's spread and avoiding severe health complications. Manual examination of blood smear samples by skilled technicians is a time-consuming aspect of the conventional malaria diagnosis toolbox. Malaria persists in many parts of the world, emphasising the urgent need for sophisticated and automated diagnostic instruments to expedite the identification of infected cells, thereby facilitating timely treatment and reducing the risk of disease transmission. This study aims to introduce a more lightweight and quicker model-but with improved accuracy-for diagnosing malaria using a YOLOv4 (You Only Look Once v. 4) deep learning object detector. METHODS: The YOLOv4 model is modified using direct layer pruning and backbone replacement. The primary objective of layer pruning is the removal and individual analysis of residual blocks within the C3, C4 and C5 (C3-C5) Res-block bodies of the backbone architecture's C3-C5 Res-block bodies. The CSP-DarkNet53 backbone is simultaneously replaced for enhanced feature extraction with a shallower ResNet50 network. The performance metrics of the models are compared and analysed. RESULTS: The modified models outperform the original YOLOv4 model. The YOLOv4-RC3_4 model with residual blocks pruned from the C3 and C4 Res-block body achieves the highest mean accuracy precision (mAP) of 90.70%. This mAP is > 9% higher than that of the original model, saving approximately 22% of the billion floating point operations (B-FLOPS) and 23 MB in size. The findings indicate that the YOLOv4-RC3_4 model also performs better, with an increase of 9.27% in detecting the infected cells upon pruning the redundant layers from the C3 Res-block bodies of the CSP-DarkeNet53 backbone. CONCLUSIONS: The results of this study highlight the use of the YOLOv4 model for detecting infected red blood cells. Pruning the residual blocks from the Res-block bodies helps to determine which Res-block bodies contribute the most and least, respectively, to the model's performance. Our method has the potential to revolutionise malaria diagnosis and pave the way for novel deep learning-based bioinformatics solutions. Developing an effective and automated process for diagnosing malaria will considerably contribute to global efforts to combat this debilitating disease. We have shown that removing undesirable residual blocks can reduce the size of the model and its computational complexity without compromising its precision.

Dengue surveillance using gravid oviposition sticky (GOS) trap and dengue non-structural 1 (NS1) antigen test in Malaysia: randomized controlled trial.

Dengue remains a major public threat and existing dengue control/surveillance programs lack sensitivity and proactivity. More efficient methods are needed. A cluster randomized controlled trial was conducted for 18 months to determine the efficacy of using a combination of gravid oviposition sticky (GOS) traps and dengue non-structural 1 (NS1) antigen for early surveillance of dengue among Aedes mosquito. Eight residential apartments were randomly assigned into intervention and control groups. GOS traps were placed at the intervention apartments weekly to trap Aedes mosquitoes and these tested for dengue NS1 antigen. When dengue-positive pool was detected, the community were notified and advised to execute protective measures. Fewer dengue cases were recorded in the intervention group than the control. Detection of NS1-positive mosquitoes was significantly associated with GOS Aedes index (rs = 0.68, P < 0.01) and occurrence of dengue cases (rs = 0.31, P < 0.01). Participants' knowledge, attitude, and practice (KAP) toward dengue control indicated significant improvement for knowledge (P < 0.01), practice (P < 0.01) and total scores (P < 0.01). Most respondents thought this surveillance method is good (81.2%) and supported its use nationwide. Thus, GOS trap and dengue NS1 antigen test can supplement the current dengue surveillance/control, in alignment with the advocated integrated vector management for reducing Aedes-borne diseases.

Survey of Dengue Knowledge and Prevention Practices Associated with Sociodemographic Status: A Cross-Sectional Study Among the Community Living in an Urban Area of Selangor, Malaysia.

The present study aimed to explore the current status of knowledge and practices of dengue prevention associated with sociodemographic status among the community living in an urban area of Selangor, Malaysia. A total of 441 participants were interviewed regarding sociodemographic status, knowledge of dengue, and self-reported prevention practices. Participants over 40 years old were more likely (odds ratio [OR] = 4.210, 95% CI = 1.652-10.733, P = 0.003) to have better dengue knowledge. Participants whose average monthly household income was more than MYR3,000 (US$715) were more likely (OR = 1.607, 95% CI = 1.059-2.438, P = 0.026) to have better practices of dengue prevention measures. The finding suggests that both government and community efforts are essential in order to continue to educate about dengue and reduce the frequency of dengue cases nationwide.

An amino acid change in nsP4 of chikungunya virus confers fitness advantage in human cell lines rather than in Aedes albopictus.

Chikungunya virus (CHIKV) has caused large-scale epidemics of fever, rash and arthritis since 2004. This unprecedented re-emergence has been associated with mutations in genes encoding structural envelope proteins, providing increased fitness in the secondary vector Aedes albopictus. In the 2008-2013 CHIKV outbreaks across Southeast Asia, an R82S mutation in non-structural protein 4 (nsP4) emerged early in Malaysia or Singapore and quickly became predominant. To determine whether this nsP4-R82S mutation provides a selective advantage in host cells, which may have contributed to the epidemic, the fitness of infectious clone-derived CHIKV with wild-type nsP4-82R and mutant nsP4-82S were compared in Ae. albopictus and human cell lines. Viral infectivity, dissemination and transmission in Ae. albopictus were not affected by the mutation when the two variants were tested separately. In competition, the nsP4-82R variant showed an advantage over nsP4-82S in dissemination to the salivary glands, but only in late infection (10 days). In human rhabdomyosarcoma (RD) and embryonic kidney (HEK-293T) cell lines coinfected at a 1 : 1 ratio, wild-type nsP4-82R virus was rapidly outcompeted by nsP4-82S virus as early as one passage (3 days). In conclusion, the nsP4-R82S mutation provides a greater selective advantage in human cells than in Ae. albopictus, which may explain its apparent natural selection during CHIKV spread in Southeast Asia. This is an unusual example of a naturally occurring mutation in a non-structural protein, which may have facilitated epidemic transmission of CHIKV.

Genetic diversity of zoonotic malaria parasites from mosquito vector and vertebrate hosts.

We explored and constructed haplotype network for simian malaria species: Plasmodium knowlesi, P. cynomolgi and P. inui aiming to understand the transmission dynamics between mosquitoes, humans and macaques. Mosquitoes were collected from villages in an area where zoonotic malaria is prevalent. PCR analysis confirmed Anopheles balabacensis as the main vector for macaque parasites, moreover nearly 60% of the mosquitoes harboured more than one Plasmodium species. Fragments of the A-type small subunit ribosomal RNA (SS rRNA) amplified from salivary gland sporozoites, and equivalent sequences obtained from GenBank were used to construct haplotype networks. The patterns were consistent with the presence of geographically distinct populations for P. inui and P. cynomolgi, and with three discrete P. knowlesi populations. This study provides a preliminary snapshot of the structure of these populations, that was insufficient to answer our aim. Thus, collection of parasites from their various hosts and over time, associated with a systematic analysis of a set of genetical loci is strongly advocated in order to obtain a clear picture of the parasite population and the flow between different hosts. This is important to devise measures that will minimise the risk of transmission to humans, because zoonotic malaria impedes malaria elimination.

Evaluation of sticky traps for adult Aedes mosquitoes in Malaysia: a potential monitoring and surveillance tool for the efficacy of control strategies.

The present study compared the performance of sticky traps in order to identify the most effective and practical trap for capturing Aedes aegypti and Aedes albopictus mosquitoes. Three phases were conducted in the study, with Phase 1 evaluating the five prototypes (Models A, B, C, D, and E) of sticky trap release-and-recapture using two groups of mosquito release numbers (five and 50) that were released in each replicate. Similarly, Phase 2 compared the performance between Model E and the classical ovitrap that had been modified (sticky ovitrap), using five and 50 mosquito release numbers. Further assessment of both traps was carried out in Phase 3, in which both traps were installed in nine sampling grids. Results from Phase 1 showed that Model E was the trap that recaptured higher numbers of mosquitoes when compared to Models A, B, C, and D. Further assessment between Model E and the modified sticky ovitrap (known as Model F) found that Model F outperformed Model E in both Phases 2 and 3. Thus, Model F was selected as the most effective and practical sticky trap, which could serve as an alternative tool for monitoring and controlling dengue vectors in Malaysia.

Mosquito Biology and Mosquito-Borne Disease Awareness Among Island Communities In Malaysia.

Mosquito-borne diseases have been increasing at an alarming rate over the past decades. In Malaysia, one finds several important mosquito-borne diseases such as Japanese encephalitis, dengue, malaria, and chikungunya. Mosquito surveillance and control programs are the most effective way of detecting and controlling mosquito-borne diseases, but these programs are less effective without an aware and well-informed general public. In 2014 we used a questionnaire to evaluate the extent of awareness of basic mosquito biology and mosquito-borne diseases in 6 villages, Kampung Masjid, Kampung Teluk Gedung, Kampung Teluk Dalam, Kampung Ujung Kelawai, Kampung Sungai Pinang Besar, and Kampung Sungai Pinang Kechil on Pangkor Island, Malaysia. A total of 1,012 individuals responded to the questionnaire, consisting of 790 Malay (78.1%), 164 Chinese (16.2%), and 58 Indian (5.7%). More than 60% (Malay = 73.7%, Chinese = 64.0%, Indian = 79.3%) of the respondents were familiar with basic mosquito biology and practiced personal protection against mosquito bites, and the association was statistically significant (P = 0.02). However, the majority of the respondents had limited knowledge on mosquito-borne diseases, and this varied significantly among the 3 ethnic groups (P = 0.0001). Our recommendations are to improve and intensify public health education outreach programs to the island residents and to encourage community participation in vector control programs.

Gastrointestinal parasites in rural dogs and cats in Selangor and Pahang states in Peninsular Malaysia.

To estimate the current prevalence of gastrointestinal (GI) parasites in dogs and cats, a total of 105 fresh faecal samples were collected from rural areas in Peninsular Malaysia. Each faecal sample was examined for the presence of GI parasites by microscopic examination after formalin-ether concentration technique and for protozoa, trichrome and Ziehl-Neelsen staining were employed. The overall prevalence of GI parasitic infection was 88.6% (95% CI = 82.5-94.7) in which 88.3% of dogs and 89.3% of cats were infected with at least one parasites species, respectively. There were 14 different GI parasites species (nematodes, cestodes and protozoa) detected, including Ancylostoma spp. (62.9%), Toxocara spp. (32.4%), Trichuris vulpis (21.0%), Spirometra spp. (9.5%), Toxascaris leonina (5.7%), Dipylidium caninum (4.8%), Ascaris spp. (2.9%), Hymenolepis diminuta (1.0%) and others. General prevalence of GI parasites showed a significant difference between helminth (84.4%) and protozoa (34.3%) infections. Monoparasitism (38.1%) was less frequent than polyparasitism (46.7%). As several of these GI parasites are recognized as zoonotic agents, the results of this investigation revealed that local populations may be exposed to a broad spectrum of zoonotic agents by means of environmental contamination with dogs and cats faeces and this information should be used to mitigate public health risks. Prevention and control measures have to be taken in order to reduce the prevalence rates especially in socioeconomically disadvantaged communities where animals live in close proximity to people, poor levels of hygiene and overcrowding together with a lack in veterinary attention and zoonotic awareness.

Plasmodium knowlesi malaria an emerging public health problem in Hulu Selangor, Selangor, Malaysia (2009-2013): epidemiologic and entomologic analysis.

BACKGROUND: While transmission of the human Plasmodium species has declined, a significant increase in Plasmodium knowlesi/Plasmodium malariae cases was reported in Hulu Selangor, Selangor, Malaysia. Thus, a study was undertaken to determine the epidemiology and the vectors involved in the transmission of knowlesi malaria. METHODS: Cases of knowlesi/malariae malaria in the Hulu Selangor district were retrospectively reviewed and analyzed from 2009 to 2013. Mosquitoes were collected from areas where cases occurred in order to determine the vectors. Leucosphyrus group of mosquitoes were genetically characterized targeting the nuclear internal transcribed spacer 2 (ITS2) and mitochondrial cytochrome c oxidase subunit I (CO1). In addition, temporal and spatial analyses were carried out for human cases and vectors. RESULTS: Of the 100 microscopy diagnosed P. knowlesi/P. malariae cases over the 5 year period in the Hulu Selangor district, there was predominance of P. knowlesi/P. malariae cases among the young adults (ages 20-39 years; 67 cases; 67%). The majority of the infected people were involved in occupations related to agriculture and forestry (51; 51%). No death was recorded in all these cases.Five hundred and thirty five mosquitoes belonging to 14 species were obtained during the study. Anopheles maculatus was the predominant species (49.5%) followed by Anopheles letifer (13.1%) and Anopheles introlatus (11.6%). Molecular and phylogenetic analysis confirmed the species of the Leucosphyrus group to be An. introlatus. In the present study, only An. introlatus was positive for oocysts. Kernel Density analysis showed that P. knowlesi hotspot areas overlapped with areas where the infected An. introlatus was discovered. This further strengthens the hypothesis that An. introlatusis is the vector for P. knowlesi in the Hulu Selangor district.Unless more information is obtained on the vectors as well as macaque involved in the transmission, it will be difficult to plan effective control strategies. The utilization of modern analytical tools such as GIS (Geographic Information System) is crucial in estimating hotspot areas for targeted control strategies. CONCLUSIONS: Anopheles introlatus has been incriminated as vector of P. knowlesi in Hulu Selangor. The cases of P. knowlesi are on the increase and further research using molecular techniques is needed.

Vertical infestation of the dengue vectors Aedes aegypti and Aedes albopictus in apartments in Kuala Lumpur, Malaysia.

Dengue is a serious public health problem in Malaysia. The aim of this study was to compare the vertical infestation of Aedes population in 2 apartments in Kuala Lumpur with different status of dengue incidence (i.e., high-dengue-incidence area and area with no reported dengue cases). The study was also conducted to assess the relationship between environmental factors such as rainfall, temperature, and humidity and Aedes population that may influence Aedes infestation. Surveillance with a mosquito larvae trapping device was conducted for 28 continuous weeks (January to July 2012) in Vista Angkasa (VA) and Inderaloka (IL) apartments located in Kuala Lumpur, Malaysia. The results indicated that both Aedes spp. could be found from ground to higher floor levels of the apartments, with Aedes aegypti being more predominant than Ae. albopictus. Data based on mixed and single breeding of Aedes spp. on different floors did not show any significant difference. Both rainfall (R3; i.e., the amount of rainfall collected during the previous 3 wk before the surveillance period began) and RH data showed significant relationship with the number of Aedes larvae collected in VA and IL. No significant difference was found between the numbers of Aedes larvae in both study areas as well as maximum and minimum temperatures. Results also indicated adaptations of Ae. aegypti to the ecosystem at each elevation of high-rise buildings, with Ae. albopictus staying inside of apartment units.