Identification of snails and parasites of medical importance via convolutional neural network: an application for human schistosomiasis

Liu, Z. Y.-C.; Chamberlin, A. J.; Shome, P.; Jones, I. J.; Riveau, G.; Ndione, R. A.; Bandagny, L.; Jouanard, N.; Van Eck, P.; Ngo, T.; Sokolow, S. H.; De Leo, G. A.

doi:10.1101/713727

bioRxiv · 10.1101/713727

Identification of snails and parasites of medical importance via convolutional neural network: an application for human schistosomiasis

Liu, Z. Y.-C.·Chamberlin, A. J.·Shome, P.·Jones, I. J.·Riveau, G.·Ndione, R. A.·Bandagny, L.·Jouanard, N.·Van Eck, P.·Ngo, T.·Sokolow, S. H.·De Leo, G. A.

View original source ↗Export metadata ↓

Abstract

Schistosomiasis is a debilitating parasitic disease infecting over 250 million people with nearly 800 million people at risk worldwide, primarily in sub-Saharan Africa. Transmission to humans involves freshwater snails as intermediate hosts, which are particularly prevalent in developing countries where dams and water resource projects have expanded freshwater snail habitat. At our study sites in the lower Senegal River Basin, we have collected more than 5,500 images of the 7 freshwater snail species (grouped into 4 categories) most frequently encountered in this aquatic ecosystem, 5 of which amplify and transmit either urinary or intestinal human schistosomiasis, with the other 2 species responsible for the transmission of less common parasitic diseases of humans and/or livestock. We have also collected over 5,100 images of 11 classes of trematodes, including human and non-human schistosomes. It takes a great deal of training and expertise to accurately classify these organisms morphologically. In recent years, deep convolutional neural networks (CNNs) have proven to be highly efficient for image recognition tasks across many object categories. Here we demonstrate classification capacity for snail and parasite images and test our models performance against 8 highly-trained human parasitologists with experience taxonomically classifying snail and parasite species from the Senegal River Basin in West Africa. We establish and train a single CNN end-to-end directly from images with only pixels and labels as inputs. Applying this state-of-the-art algorithm, we are able to classify images of 4 snail categories with 99.64% accuracy and images of 11 parasite categories with 88.14% accuracy, which rivals highly-trained human parasitologists. The trained algorithm could next be deployed to mobile devices for use in remote field settings by local technicians, and significantly improve monitoring snail and parasite occurrence in the field for disease control purposes.\n\nAuthor SummarySchistosomiasis is a neglected tropical disease (NTD) infecting over 250 million people worldwide. The current approach to mitigate this disease in endemic regions is community or school-based mass drug administration. However, parasites are primarily transmitted through environmental reservoirs where freshwater snails serve as intermediate hosts. People use the contaminated water sources for their daily tasks and get re-infected after drug treatment. Therefore, drug administration alone is not effective for schistosomiasis control in such high transmission settings. Recent studies show that snail population control is essential to reduce disease transmission risks. To discern between parasitic worms of humans in snails and those of other non-human species is a necessary step to precisely quantify transmission risk for human schistosomiasis. However, it takes a great deal of expertise to train lab and field technicians to accurately classify snail and parasite species. Artificial intelligence (AI)-powered computer vision algorithms have recently proven to be highly efficient for image recognition tasks. We have collected thousands of snail and parasite images in Senegal during 2015-2018. Using these images as training data, we developed an AI model that classifies images of 4 snail categories and 11 parasite categories with, in some cases, higher accuracy than well-trained human parasitologists. This model could next be deployed to mobile devices for use in remote field settings to support local technicians to identify transmission hotspots and target control.

Explore related subjects

bioinformatics Preprint

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Liu, Z. Y.-C., Chamberlin, A. J., Shome, P., Jones, I. J., Riveau, G., Ndione, R. A., Bandagny, L., Jouanard, N., Van Eck, P., Ngo, T., Sokolow, S. H., De Leo, G. A.. 2019-07-24. Identification of snails and parasites of medical importance via convolutional neural network: an application for human schistosomiasis. https://doi.org/10.1101/713727

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Contrastive Learning Improves Generalization on M2OR Olfactory Binding Dataset

Olfaction-based biosensors stand to have profound impacts in fields such as medical diagnostics, national security, and drug discovery. A promising mode of vapor phase biodetection mimics volatile organic compound-olfactory receptor (VOC-OR) interactions with target-specific protein or peptide probes; however, lack of experimental coverage remains a core bottleneck. Current computational approaches aim to improve prediction of VOC-OR binding to reduce experimental burden, and current machine learning approaches fail to demonstrate a level of generalization that addresses current engineering challenges. Here, using the Molecular to Olfactory Dataset (M2OR) and Insect Olfactory Binding Protein Database (iOBPdb) datasets, data partitioning methods and their performance are evaluated on a neural network and contrastive learning framework on such VOC-OR binding datasets, varying: model training, architecture, and structure of the input data distribution. It was found that holding out certain chemical or proteomic motifs from the training data diminished model performance with a two layer neural network model. The model was able to generalize to state of the art performance (0.88 AUCROC) on every data split by applying a contrastive learning objective defined by shared molecular constituents and a stochastic sampler. This suggests the contrastive objective provides a strong inductive bias for neural networks to predict novel VOC-OR pairs. These findings highlight that model performance on pairwise biological interaction datasets can be inflated through identity leakage from constituent overlap, and emphasize the need for more robust evaluation protocols. Furthermore, we show contrastive regularization mitigates dataset degeneracy through enforcement of compatibility in the models latent geometry that may extend to other binder target-datasets where identity leakage confounds evaluation.

bioinformatics↗

bioRxiv2026

Plate design confounds donor variation and model evaluation in single-cell perturbation screens

Single-cell screens of immune cells from many donors are used to study why individuals respond differently and to benchmark response models. Each donor-perturbation combination typically occupies one well, controls are grouped, and donors share one layout, so technical effects can look biological. Using a 90-cytokine screen, two compound screens, four cohorts and simulations, we show that well effects explain most replicate-control variation in abundant cell types, controls carry a shared offset, and identical control wells differ by hundreds of differentially expressed genes at a standard threshold, about as many as two receptor-negative proteins do. Negative controls, design-aware estimators and perturbation-specific evaluation change the interpretation of four recent analyses. A deep-learning model that scores below a perturbation-agnostic predictor under standard evaluation outperforms a simple baseline on perturbation-specific differences in two of three cell types (significantly in one with stricter controls), but predicts less well the component shared with inert wells.

bioinformatics↗

bioRxiv2026

LinearRaccess: linear-time approximation of RNA accessibility without base-pair span limits

Motivation: RNA accessibility is the probability that every nucleotide in a contiguous interval is simultaneously unpaired; it is widely used to predict binding sites for RNA-binding proteins and other RNA-RNA interactions, often in the form of the opening energy, an equivalent measure on a logarithmic scale. Existing tools restrict the maximum base-pair span, discarding long-range base pairs that matter biologically. Others sample structures, introducing statistical variance, or compute accessibility exactly without a span limit, at cubic time cost. No existing method computes the accessibility of windows longer than one nucleotide deterministically in linear time while preserving long-range base pairs. We present LinearRaccess, which computes accessibility directly in linear time via beam-pruned inside-outside dynamic programming with a difference-array aggregation step, without restricting base-pair span. Results: For a sequence of length N and a fixed beam width b, LinearRaccess runs in O(N) time and O(Nb) memory. Against exact accessibility on 12,872 bpRNA-1m(90) sequences of 100-1,000 nt, its median opening-energy error is below 0.001 kcal/mol (0.14 kcal/mol for 700-1,000 nt), and it distinguishes windows of very small accessibility, including 35-nt windows, below the resolution of sampling-based estimates. LinearRaccess analyzes the full 29,903-nt SARS-CoV-2 genome in 419 seconds. Conclusion: LinearRaccess is the first method to compute the accessibility of windows longer than one nucleotide deterministically (without sampling) in linear time without restricting base-pair span. Availability: https://github.com/TakumiOtagaki/LinearRaccess, archived at Zenodo (https://doi.org/10.5281/zenodo.23598502).

bioinformatics↗