Search for: All records

Through the COVID-19 pandemic, SARS-CoV-2 has gained and lost multiple mutations in novel or unexpected combinations. Predicting how complex mutations affect COVID-19 disease severity is critical in planning public health responses as the virus continues to evolve. This paper presents a novel computational framework to complement conventional lineage classification and applies it to predict the severe disease potential of viral genetic variation. The transformer-based neural network model architecture has additional layers that provide sample embeddings and sequence-wide attention for interpretation and visualization. First, training a model to predict SARS-CoV-2 taxonomy validates the architecture’s interpretability. Second, an interpretable predictive model of disease severity is trained on spike protein sequence and patient metadata from GISAID. Confounding effects of changing patient demographics, increasing vaccination rates, and improving treatment over time are addressed by including demographics and case date as independent input to the neural network model. The resulting model can be interpreted to identify potentially significant virus mutations and proves to be a robust predctive tool. Although trained on sequence data obtained entirely before the availability of empirical data for Omicron, the model can predict the Omicron’s reduced risk of severe disease, in accord with epidemiological and experimental data. 

Efficiently and accurately identifying which microbes are present in a biological sample is important to medicine and biology. For example, in medicine, microbe identification allows doctors to better diagnose diseases. Two questions are essential to metagenomic analysis (the analysis of a random sampling of DNA in a patient/environment sample): How to accurately identify the microbes in samples and how to efficiently update the taxonomic classifier as new microbe genomes are sequenced and added to the reference database. To investigate how classifiers change as they train on more knowledge, we made sub-databases composed of genomes that existed in past years that served as “snapshots in time” (1999–2020) of the NCBI reference genome database. We evaluated two classification methods, Kraken 2 and CLARK with these snapshots using a real, experimental metagenomic sample from a human gut. This allowed us to measure how much of a real sample could confidently classify using these methods and as the database grows. Despite not knowing the ground truth, we could measure the concordance between methods and between years of the database within each method using a Bray-Curtis distance. In addition, we also recorded the training times of the classifiers for each snapshot. For all data for Kraken 2, we observed that as more genomes were added, more microbes from the sample were classified. CLARK had a similar trend, but in the final year, this trend reversed with the microbial variation and less unique k-mers. Also, both classifiers, while having different ways of training, generally are linear in time - but Kraken 2 has a significantly lower slope in scaling to more data. 

Recurrent neural networks with memory and attention mechanisms are widely used in natural language processing because they can capture short and long term sequential information for diverse tasks. We propose an integrated deep learning model for microbial DNA sequence data, which exploits convolutional neural networks, recurrent neural networks, and attention mechanisms to predict taxonomic classifications and sample-associated attributes, such as the relationship between the microbiome and host phenotype, on the read/sequence level. In this paper, we develop this novel deep learning approach and evaluate its application to amplicon sequences. We apply our approach to short DNA reads and full sequences of 16S ribosomal RNA (rRNA) marker genes, which identify the heterogeneity of a microbial community sample. We demonstrate that our implementation of a novel attention-based deep network architecture, Read2Pheno , achieves read-level phenotypic prediction. Training Read2Pheno models will encode sequences (reads) into dense, meaningful representations: learned embedded vectors output from the intermediate layer of the network model, which can provide biological insight when visualized. The attention layer of Read2Pheno models can also automatically identify nucleotide regions in reads/sequences which are particularly informative for classification. As such, this novel approach can avoid pre/post-processing and manual interpretation required with conventional approaches to microbiome sequence classification. We further show, as proof-of-concept, that aggregating read-level information can robustly predict microbial community properties, host phenotype, and taxonomic classification, with performance at least comparable to conventional approaches. An implementation of the attention-based deep learning network is available at https://github.com/EESI/sequence_attention (a python package) and https://github.com/EESI/seq2att (a command line tool). 

It is critical for biological studies to annotate amino acid sequences and understand how proteins function. Protein function is important to medical research in the health industry (e.g., drug discovery). With the advancement of deep learning, accurate protein annotation models have been developed for alignment free protein annotation. In this paper, we develop a deep learning model with an attention mechanism that can predict Gene Ontology labels given a protein sequence input. We believe this model can produce accurate predictions as well as maintain good interpretability. We further show how the model can be interpreted by examining and visualizing the intermediate layer output in our deep neural network. 

Abstract Background It is a computational challenge for current metagenomic classifiers to keep up with the pace of training data generated from genome sequencing projects, such as the exponentially-growing NCBI RefSeq bacterial genome database. When new reference sequences are added to training data, statically trained classifiers must be rerun on all data, resulting in a highly inefficient process. The rich literature of “incremental learning” addresses the need to update an existing classifier to accommodate new data without sacrificing much accuracy compared to retraining the classifier with all data. Results We demonstrate how classification improves over time by incrementally training a classifier on progressive RefSeq snapshots and testing it on: (a) all known current genomes (as a ground truth set) and (b) a real experimental metagenomic gut sample. We demonstrate that as a classifier model’s knowledge of genomes grows, classification accuracy increases. The proof-of-concept naïve Bayes implementation, when updated yearly, now runs in 1/4 t h of the non-incremental time with no accuracy loss. Conclusions It is evident that classification improves by having the most current knowledge at its disposal. Therefore, it is of utmost importance to make classifiers computationally tractable to keep up with the data deluge. The incremental learning classifier can be efficiently updated without the cost of reprocessing nor the access to the existing database and therefore save storage as well as computation resources. 

Viral subtyping can facilitate visualization and modeling of the geographic distribution and temporal dynamics of disease spread. Understanding the virus's evolution spatiotemporally can help forensic strategies. We have identified mutation variation within SARS-CoV-2 sequences via an entropy measure followed by frequency analysis. These signatures, Informative Subtype Markers (ISMs), define a compact set of nucleotide sites that characterize the most variable (and thus most informative) positions in the viral genomes sequenced from different individuals. Using these ISMs, we show that we can use them for a variety of downstream analyses, such as comparing countries' subtype compositions. We present association graphs as a visualization tool to connect different ISMs based on their co-occurrence across different individuals. In particular, we investigate dominant ISMs for different locations, across different factors such as gender and age. 

Machine learning algorithms can learn mechanisms of antimicrobial resistance from the data of DNA sequence without any a priori information. Interpreting a trained machine learning algorithm can be exploited for validating the model and obtaining new information about resistance mechanisms. Different feature extraction methods, such as SNP calling and counting nucleotide k-mers have been proposed for presenting DNA sequences to the model. However, there are trade-offs between interpretability, computational complexity and accuracy for different feature extraction methods. In this study, we have proposed a new feature extraction method, counting amino acid k-mers or oligopeptides, which provides easier model interpretation compared to counting nucleotide k-mers and reaches the same or even better accuracy in comparison with different methods. Additionally, we have trained machine learning algorithms using different feature extraction methods and compared the results in terms of accuracy, model interpretability and computational complexity. We have built a new feature selection pipeline for extraction of important features so that new AMR determinants can be discovered by analyzing these features. This pipeline allows the construction of models that only use a small number of features and can predict resistance accurately. 

Abstract Evaluating metagenomic software is key for optimizing metagenome interpretation and focus of the Initiative for the Critical Assessment of Metagenome Interpretation (CAMI). The CAMI II challenge engaged the community to assess methods on realistic and complex datasets with long- and short-read sequences, created computationally from around 1,700 new and known genomes, as well as 600 new plasmids and viruses. Here we analyze 5,002 results by 76 program versions. Substantial improvements were seen in assembly, some due to long-read data. Related strains still were challenging for assembly and genome recovery through binning, as was assembly quality for the latter. Profilers markedly matured, with taxon profilers and binners excelling at higher bacterial ranks, but underperforming for viruses and Archaea. Clinical pathogen detection results revealed a need to improve reproducibility. Runtime and memory usage analyses identified efficient programs, including top performers with other metrics. The results identify challenges and guide researchers in selecting methods for analyses.