Temporal deep learning architecture for prediction of COVID-19 cases in India.

To combat the recent coronavirus disease 2019 (COVID-19), academician and clinician are in search of new approaches to predict the COVID-19 outbreak dynamic trends that may slow down or stop the pandemic. Epidemiological models like Susceptible-Infected-Recovered (SIR) and its variants are helpful to understand the dynamics trend of pandemic that may be used in decision making to optimize possible controls from the infectious disease. But these epidemiological models based on mathematical assumptions may not predict the real pandemic situation. Recently the new machine learning approaches are being used to understand the dynamic trend of COVID-19 spread. In this paper, we designed the recurrent and convolutional neural network models: vanilla LSTM, stacked LSTM, ED_LSTM, BiLSTM, CNN, and hybrid CNN+LSTM model to capture the complex trend of COVID-19 outbreak and perform the forecasting of COVID-19 daily confirmed cases of 7, 14, 21 days for India and its four most affected states (Maharashtra, Kerala, Karnataka, and Tamil Nadu). The root mean square error (RMSE) and mean absolute percentage error (MAPE) evaluation metric are computed on the testing data to demonstrate the relative performance of these models. The results show that the stacked LSTM and hybrid CNN+LSTM models perform best relative to other models.

Pattern of genomic variation in SARS-CoV-2 (COVID-19) suggests restricted nonrandom changes: Analysis using Shewhart control charts.

SARS-CoV-2 is a member of the Coronavirus family which recently originated from the Wuhan province of China and spread very rapidly through the world infecting more than 4 million people. In the past, other Coronaviruses have also been found to cause human infection, but not as widespread as COVID-19. Since Coronavirus sequences constantly change due to mutation and recombination, it is important to understand the pattern of changes and likely path the virus can take in the future. In this study, we have used the Shewhart control chart to identify and analyze hypervariable (hotspots) and hypovariable (coldspots) regions of the virus. Our analysis shows that SARS-CoV-2 has changed in a few regions of the genome. Analysis of SARS-CoV-1 and MERS sequences suggests that over time, mutations start accumulating in different regions and most likely SARS-CoV-2 may also follow a similar path. The results suggest a possible emergence of modified viruses over some time.

Comprehensive Needs Assessment to Ensure Appropriate Rehabilitation Training for Community-Based Workers and Caregivers in India.

IMPORTANCE: Rehabilitation researchers and occupational therapy practitioners frequently conduct projects that involve the provision of services or the dissemination of knowledge in low- and middle-income countries (LMICs); however, the needs assessment process underlying these transcultural projects is often underreported. OBJECTIVE: To identify a comprehensive set of feeding-related training needs that reflect the culture, resources, and service delivery practices in the West Bengal region of India and that will be used to develop video-based training modules that are contextually relevant and support decentralized training of community-based rehabilitation workers and caregivers of children with cerebral palsy (CP). DESIGN: Mixed methods design using focus groups and observations. SETTING: Community setting in rural, suburban, and urban locations in West Bengal, India. PARTICIPANTS: Caregivers of children with CP and community-based rehabilitation workers with experience working with children with CP. OUTCOMES AND MEASURES: Focus groups and in-home mealtime observations. RESULTS: Several findings suggest training priorities that were not anticipated. Caregivers commonly fed children by means of hand feeding while sitting on the floor. Children were fed a limited variation of food consistencies, often in positions of neck hyperextension. Water was frequently used to stimulate swallowing. Caregivers struggled to help children with oral hygiene practices after meals. CONCLUSIONS AND RELEVANCE: A thorough needs assessment yielded insights that can favorably alter the trajectory of transcultural service and research projects. For occupational therapists who provide advice on practice, continuing education, or research in a transcultural project, the findings underscore the value of engaging local stakeholders and using firsthand observation. WHAT THIS ARTICLE ADDS: The rigorous, multistakeholder needs assessment process described herein can serve as a road map for researchers and occupational therapy practitioners conducting transcultural projects in LMICs.

Stochastic method to control Mycobacterium tuberculosis epidemic.

We study the origin of TB (tuberculosis) epidemic and complex distributions of various populations of TB infection within the stochastic framework. The stochastic nature of this disease infection could be linked to the stochastic behaviour at genome level which is exhibited in SNP (single nucleotide polymorphism) distributions of experimentally identified hotspot driver genes. Our results show the emergence of random clusters, and well-defined discrete domains of the respective species populations in the model driven by demographic stochasticity and intrinsic complex species interaction. The multifractal analysis of the time series of the species populations indicate that TB epidemic could be mainly caused by contact communication and is directional. We propose that any TB epidemic may have high chance of approximately periodic recurrence and can be controlled by optimizing some of the parameters involved in the system modelling.

Complex multifractal nature in Mycobacterium tuberculosis genome.

The mutifractal and long range correlation (C(r)) properties of strings, such as nucleotide sequence can be a useful parameter for identification of underlying patterns and variations. In this study C(r) and multifractal singularity function f(α) have been used to study variations in the genomes of a pathogenic bacteria Mycobacterium tuberculosis. Genomic sequences of M. tuberculosis isolates displayed significant variations in C(r) and f(α) reflecting inherent differences in sequences among isolates. M. tuberculosis isolates can be categorised into different subgroups based on sensitivity to drugs, these are DS (drug sensitive isolates), MDR (multi-drug resistant isolates) and XDR (extremely drug resistant isolates). C(r) follows significantly different scaling rules in different subgroups of isolates, but all the isolates follow one parameter scaling law. The richness in complexity of each subgroup can be quantified by the measures of multifractal parameters displaying a pattern in which XDR isolates have highest value and lowest for drug sensitive isolates. Therefore C(r) and multifractal functions can be useful parameters for analysis of genomic sequences.

Analysis of IS6110 insertion sites provide a glimpse into genome evolution of Mycobacterium tuberculosis.

Insertion sequence (IS) 6110 is found at multiple sites in the Mycobacterium tuberculosis genome and displays a high degree of polymorphism with respect to copy number and insertion sites. Therefore, IS6110 is considered to be a useful molecular marker for diagnosis and strain typing of M. tuberculosis. Generally IS6110 elements are identified using experimental methods, useful for analysis of a limited number of isolates. Since short read genome sequences generated using next-generation sequencing (NGS) platforms are available for a large number of isolates, a computational pipeline for identification of IS6110 elements from these datasets was developed. This study shows results from analysis of NGS data of 1377 M. tuberculosis isolates. These isolates represent all seven major global lineages of M. tuberculosis. Lineage specific copy number patterns and preferential insertion regions were observed. Intra-lineage differences were further analyzed for identifying spoligotype specific variations. Copy number distribution and preferential locations of IS6110 in different lineages imply independent evolution of IS6110, governed mainly through ancestral insertion, fitness (gene truncation, promoter activity) and recombinational loss of some copies. A phylogenetic tree based on IS6110 insertion data of different isolates was constructed in order to understand genome level variations of different markers across different lineages.