Background Transfer learning is a form of machine learning where a pre-trained model trained on a specific task is reused as a starting point and tailored to another task in a different dataset. While transfer learning has garnered considerable attention in medical image analysis, its use for clinical non-image data is not well studied. Therefore, the objective of this scoping review was to explore the use of transfer learning for non-image data in the clinical literature. Methods and Findings We systematically searched medical databases (PubMed, EMBASE, CINAHL) for peer-reviewed clinical studies that used transfer learning on human non-image data. We included 83 studies in the review. More than half of the studies (63%) were published within 12 months of the search. Transfer learning was most often applied to time series data (61%), followed by tabular data (18%), audio (12%) and text (8%). Thirty-three (40%) studies applied an image-based model to non-image data after transforming data into images (e.g. spectrograms). Twenty-nine (35%) studies did not have any authors with a health-related affiliation. Many studies used publicly available datasets (66%) and models (49%), but fewer shared their code (27%). Conclusions In this scoping review, we have described current trends in the use of transfer learning for non-image data in the clinical literature. We found that the use of transfer learning has grown rapidly within the last few years. We have identified studies and demonstrated the potential of transfer learning in clinical research in a wide range of medical specialties. More interdisciplinary collaborations and the wider adaption of reproducible research principles are needed to increase the impact of transfer learning in clinical research.

https://medrxiv.org/content/10.1101/2021.10.01.21264290v1.full.pdf

Transfer learning for non-image data in clinical research: a scoping review

Background Transfer learning is a form of machine learning where a pre-trained model trained on a specific task is reused as a starting point and tailored to another task in a different dataset. While transfer learning has garnered considerable attention in medical image analysis, its use for clinical non-image data is not well studied. Therefore, the objective of this scoping review was to explore the use of transfer learning for non-image data in the clinical literature. Methods and findings We systematically searched medical databases (PubMed, EMBASE, CINAHL) for peer-reviewed clinical studies that used transfer learning on human non-image data. We included 83 studies in the review. More than half of the studies (63%) were published within 12 months of the search. Transfer learning was most often applied to time series data (61%), followed by tabular data (18%), audio (12%) and text (8%). Thirty-three (40%) studies applied an image-based model to non-image data after transforming data into images (e.g. spectrograms). Twenty-nine (35%) studies did not have any authors with a health-related affiliation. Many studies used publicly available datasets (66%) and models (49%), but fewer shared their code (27%). Conclusions In this scoping review, we have described current trends in the use of transfer learning for non-image data in the clinical literature. We found that the use of transfer learning has grown rapidly within the last few years. We have identified studies and demonstrated the potential of transfer learning in clinical research in a wide range of medical specialties. More interdisciplinary collaborations and the wider adaption of reproducible research principles are needed to increase the impact of transfer learning in clinical research. 

/pdf/transfer-learning-for-non-image-data-in-clinical-research-a-ulhtrht7.pdf

Transfer learning for non-image data in clinical research: A scoping review

Coronavirus disease 2019 (COVID-19) is as an emerging infectious disease (EID) that has caused the worst public health catastrophe of the 21st century thus far. In terms of impact, the COVID-19 pandemic is second only to the Spanish Flu pandemic of 1918 in modern world history. As of 7 September 2021, there have been 220 million confirmed cases of COVID-19 and more than 4.5 million deaths. EIDs pose serious public health and socio-economic risks, and 70% of EIDs originate from wildlife. Preventing development of EIDs such as COVID-19 is a pressing concern. Here, taking the COVID-19 pandemic as an example, we illustrate the disastrous effects of EIDs and assess their emergence and evolution from a One Health perspective. We propose a One Health strategy, centered on 'moving the gates forward', for EID prevention and control at the human-animal-environment interface. This strategy may be instructive and provide early warnings of EIDs in the future. 

A One Health strategy for emerging infectious diseases based on the COVID-19 outbreak

Effective prediction of drug – Target interaction on HIV using deep graph neural networks

Whereas accelerated attention beclouded early stages of the coronavirus spread, knowledge of actual pathogenicity and origin of possible sub-strains remained unclear. By harvesting the Global initiative on Sharing All Influenza Data (GISAID) database ( https://www.gisaid.org/ ), between December 2019 and January 15, 2021, a total of 8864 human SARS-CoV-2 complete genome sequences processed by gender, across 6 continents (88 countries) of the world, Antarctica exempt, were analyzed. We hypothesized that data speak for itself and can discern true and explainable patterns of the disease. Identical genome diversity and pattern correlates analysis performed using a hybrid of biotechnology and machine learning methods corroborate the emergence of inter- and intra- SARS-CoV-2 sub-strains transmission and sustain an increase in sub-strains within the various continents, with nucleotide mutations dynamically varying between individuals in close association with the virus as it adapts to its host/environment. Interestingly, some viral sub-strain patterns progressively transformed into new sub-strain clusters indicating varying amino acid, and strong nucleotide association derived from same lineage. A novel cognitive approach to knowledge mining helped the discovery of transmission routes and seamless contact tracing protocol. Our classification results were better than state-of-the-art methods, indicating a more robust system for predicting emerging or new viral sub-strain(s). The results therefore offer explanations for the growing concerns about the virus and its next wave(s). A future direction of this work is a defuzzification of confusable pattern clusters for precise intra-country SARS-CoV-2 sub-strains analytics.

/pdf/a-hybrid-computational-framework-for-intelligent-inter-15wmdtz1bg.pdf

A hybrid computational framework for intelligent inter-continent SARS-CoV-2 sub-strains characterization and prediction.

Background: The COVID-19 pandemic has ravaged economies, health systems, and lives globally. Concerns surrounding near total economic collapse, loss of livelihood and emotional complications ensuing from lockdowns and commercial inactivity, resulted in governments loosening economic restrictions. These concerns were further exacerbated by the absence of vaccines and drugs to combat the disease, with the fear that the next wave of the pandemic would be more fatal. Consequently, integrating disease surveillance mechanism into public healthcare systems is gaining traction, to reduce the spread of community and cross-border infections and offer informed medical decisions. Methods: Publicly available datasets of coronavirus cases around the globe deposited between December, 2019 and March 15, 2021 were retrieved from GISAID EpiFluTM and processed. Also retrieved from GISAID were data on the different SARS-CoV-2 variant types since inception of the pandemic. Results: Epidemiological analysis offered interesting statistics for understanding the demography of SARS-CoV-2 and helped the elucidation of local and foreign transmission through a history of contact travels. Results of genome pattern visualization and cognitive knowledge mining revealed the emergence of high intra-country viral sub-strains with localized transmission routes traceable to immediate countries, for enhanced contact tracing protocol. Variant surveillance analysis indicates increased need for continuous monitoring of SARS-CoV-2 variants.  A collaborative Internet of Health Things (IoHT) framework was finally proposed to impact the public health system, for robust and intelligent support for modelling, characterizing, diagnosing and real-time contact tracing of infectious diseases. Conclusions: Localizing healthcare disease surveillance is crucial in emerging disease situations and will support real-time/updated disease case definitions for suspected and probable cases. The IoHT framework proposed in this paper will assist early syndromic assessments of emerging infectious diseases and support healthcare/medical countermeasures as well as useful strategies for making informed policy decisions to drive a cost effective, smart healthcare system.

/pdf/sars-cov-2-genome-datasets-analytics-for-informed-infectious-4w95clyksz.pdf

SARS-CoV-2 genome datasets analytics for informed infectious disease surveillance

In middle- and low-income countries where higher nosocomial infection rates have been reported, approximately 5% to 10% of hospitalized patients have some infection acquired after admission. Recent studies suggest that contaminated environmental surfaces may play a major role in the transmission of nosocomial infections. Therefore, this study presents an e-Medical Assistant Tool (Vitex), which is a mobile device that disinfects wards of 100 square feet in a single cycle which can be increased since the device is mobile by using powerful U.V rays of 222nm that can be used in occupied rooms without adverse effects on human health. It also employs artificial intelligence, big data, and machine learning to improve patient care and practitioner assistance. Unlike existing similar devices, our innovation disinfects the ward, and facilitates healthcare provision via remote patient consultation and diagnosis; thus, bringing care nearer to the patient.

Tackling Occupational and Nosocomial Infection using Vitex-Medical Assistant Tool

BACKGROUND
Effective global antiretroviral vaccines and therapeutic strategies depend on the diversity, evolution, and epidemiology of their various strains as well as their transmission and pathogenesis. Most viral disease-causing particles are clustered into a taxonomy of subtypes to suggest pointers toward nucleotide-specific vaccines or therapeutic applications of clinical significance sufficient for sequence-specific diagnosis and homologous viral studies. These are very useful to formulate predictors to induce cross-resistance to some retroviral control drugs being used across study areas.


OBJECTIVE
This research proposed a collaborative framework of hybridized (Machine Learning and Natural Language Processing) techniques to discover hidden genome patterns and feature predictors, for HIV-1 genome sequences mining.


METHOD
630 human HIV-1 genome sequences above 8500 bps were excavated from the National Center for Biotechnology Information (NCBI) database (https://www.ncbi.nlm.nih.gov) for 21 countries across different continents, Antarctica exempt. These sequences were transformed and learned using a self-organizing map (SOM). To discriminate emerging/new sub-strain(s), the HIV-1 reference genome was included as part of the input isolates/samples during the training. After training the SOM, component planes defining pattern clusters of the input datasets were generated, for cognitive knowledge mining and subsequent labelling of the datasets. Additional genome features including dinucleotide transmission recurrences, codon recurrences, and mutation recurrences, were finally extracted from the raw genomes to construct output classification targets for supervised learning.


RESULTS
SOM training explains the inherent pattern diversity of HIV-1 genomes as well as inter- and intra-country transmissions in which mobility might play an active role, as corroborated by literature. Nine sub-strains were discovered after disassembling the SOM correlation hunting matrix space attributed to disparate clusters. Cognitive knowledge mining separated similar pattern clusters bounded by a certain degree of correlation range, discovered by the SOM. A Kruskal-Wallis rank-sum test and Wilcoxon rank-sum test showed statistically significant variations in dinucleotide, codon, and mutation patterns.


CONCLUSION
Results of the discovered sub-strains and response clusters visualizations corroborate existing literature, with significant haplotype variations. The proposed framework would assist in the development of decision support systems for easy contact tracing, infectious disease surveillance, and studying the progressive evolution of the reference HIV-1 genome.

Collaborative Mining of Whole Genome Sequences for Intelligent HIV-1 Sub-Strain(s) Discovery.


 BackgroundThe increased number of accessible genomes has prompted large-scale comparative studies for decerning evolutionary knowledge of infectious diseases, but challenges such as non-availability of close reference sequence(s), incompletely assembled or large number of genomes, preclude real time multiple sequence alignment and sub-strain(s) discovery. This paper introduces a cooperatively inspired open-source framework, for intelligent mining of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) genomes. We situate this study within the African context, to drive advancement on state-of-the-art, towards intelligent infectious disease characterization and prediction. The outcome is an enriched Knowledge Base, sufficient to provide deep understanding of the viral sub-strains’ identification problem. We also open investigation by gender, which to the best of our knowledge has been ignored in related research. Data for the study came from the Global Initiative on Sharing All Influenza Data database (https://gisaid.org) and processed for precise discovery of viral sub-strains transmission between and within African countries. To localize the transmission route(s) of each isolate excavated and provide appropriate links to similar isolate strain(s), a cognitive solution was imposed on the genome expression patterns discovered by unsupervised self-organizing map (SOM) component planes visualization. The Freidman-Nemenyi’s test was finally performed to validate our claim.ResultsEvidence of inter- and intra-genome diversity was noticed. While some isolates (or genomes) clustered differently, implying different evolutionary source (or high-diversity), others clustered closely together, indicating similar evolutionary source (or less-diversity). SOM component planes analysis revealed multiple sub-strains patterns, strongly suggesting local or intra-community and country to country transmissions. Cognitive maps of both male and female isolates revealed multiple transmission routes. Statistical results indicate significant difference between the various isolate groups at the 0.05 level of significance.ConclusionThe proposed framework offers explanations to SARS-CoV-2 diversity and provides real time identification to disease transmission routes, as well as rapid decision support for facilitating inter- and intra-country contact tracing of infected case(s). Intermediate data produced in this paper are helpful to enrich the genome datasets for intelligent characterization and prediction of COVID-19 and related pandemics, as well as the construction of intelligent device for accurate infectious disease monitoring.

/pdf/real-time-viral-sub-strains-discovery-in-emerging-infectious-4eesjyazvt.pdf

Mercy E. Edoho

Papers

SARS-CoV-2 genome datasets analytics for informed infectious disease surveillance

Tackling Occupational and Nosocomial Infection using Vitex-Medical Assistant Tool

Collaborative Mining of Whole Genome Sequences for Intelligent HIV-1 Sub-Strain(s) Discovery.

Real Time Viral Sub-Strains Discovery in Emerging Infectious Disease Situation – The African Perspective