Improvements in viral gene annotation using large language models and soft alignments

BACKGROUND

The annotation of protein sequences in public databases has long posed a challenge in molecular biology. This issue is particularly acute for viral proteins, which demonstrate limited homology to known proteins when using alignment, k-mer, or profile-based homology search approaches. A novel methodology employing Large Language Models (LLMs) addresses this methodological challenge by annotating protein sequences based on embeddings.

RESULTS

Central to our contribution is the soft alignment algorithm, drawing from traditional protein alignment but leveraging embedding similarity at the amino acid level to bypass the need for conventional scoring matrices. This method not only surpasses pooled embedding-based models in efficiency but also in interpretability, enabling users to easily trace homologous amino acids and delve deeper into the alignments. Far from being a black box, our approach provides transparent, BLAST-like alignment visualizations, combining traditional biological research with AI advancements to elevate protein annotation through embedding-based analysis while ensuring interpretability. Tests using the Virus Orthologous Groups and ViralZone protein databases indicated that the novel soft alignment approach recognized and annotated sequences that both blastp and pooling-based methods, which are commonly used for sequence annotation, failed to detect.

CONCLUSION

The embeddings approach shows the great potential of LLMs for enhancing protein sequence annotation, especially in viral genomics. These findings present a promising avenue for more efficient and accurate protein function inference in molecular biology.

Leveraging deep contrastive learning for semantic interaction

The semantic interaction process seeks to elicit a user's mental model as they interact with and query visualizations during a sense-making activity. Semantic interaction enables the development of computational models that capture user intent and anticipate user actions. Deep learning is proving to be highly effective for learning complex functions and is, therefore, a compelling tool for encoding a user's mental model. In this paper, we show that deep contrastive learning significantly enhances semantic interaction in visual analytics systems. Our approach does so by allowing users to explore alternative arrangements of their data while simultaneously training a parametric algorithm to learn their evolving mental model. As an example of the efficacy of our approach, we deployed our model in Z-Explorer, a visual analytics extension to the widely used Zotero document management system. The user study demonstrates that this flexible approach effectively captures users' mental data models without explicit hyperparameter tuning or even requiring prior machine learning expertise.

CoCoNet: an efficient deep learning tool for viral metagenome binning

MOTIVATION

Metagenomic approaches hold the potential to characterize microbial communities and unravel the intricate link between the microbiome and biological processes. Assembly is one of the most critical steps in metagenomics experiments. It consists of transforming overlapping DNA sequencing reads into sufficiently accurate representations of the community's genomes. This process is computationally difficult and commonly results in genomes fragmented across many contigs. Computational binning methods are used to mitigate fragmentation by partitioning contigs based on their sequence composition, abundance or chromosome organization into bins representing the community's genomes. Existing binning methods have been principally tuned for bacterial genomes and do not perform favorably on viral metagenomes.

RESULTS

We propose Composition and Coverage Network (CoCoNet), a new binning method for viral metagenomes that leverages the flexibility and the effectiveness of deep learning to model the co-occurrence of contigs belonging to the same viral genome and provide a rigorous framework for binning viral contigs. Our results show that CoCoNet substantially outperforms existing binning methods on viral datasets.

AVAILABILITY AND IMPLEMENTATION

CoCoNet was implemented in Python and is available for download on PyPi (https://pypi.org/). The source code is hosted on GitHub at https://github.com/Puumanamana/CoCoNet and the documentation is available at https://coconet.readthedocs.io/en/latest/index.html. CoCoNet does not require extensive resources to run. For example, binning 100k contigs took about 4 h on 10 Intel CPU Cores (2.4 GHz), with a memory peak at 27 GB (see Supplementary Fig. S9). To process a large dataset, CoCoNet may need to be run on a high RAM capacity server. Such servers are typically available in high-performance or cloud computing settings.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Transcriptional regulation of prolactin in a euryhaline teleost: Characterisation of gene promoters through in silico and transcriptome analyses

The sensitivity of prolactin (Prl) cells of the Mozambique tilapia (Oreochromis mossambicus) pituitary to variations in extracellular osmolality enables investigations into how osmoreception underlies patterns of hormone secretion. Through the actions of their main secretory products, Prl cells play a key role in supporting hydromineral balance of fishes by controlling the major osmoregulatory organs (ie, gill, intestine and kidney). The release of Prl from isolated cells of the rostral pars distalis (RPD) occurs in direct response to physiologically relevant reductions in extracellular osmolality. Although the particular signal transduction pathways that link osmotic conditions to Prl secretion have been identified, the processes that underlie hyposmotic induction of prl gene expression remain unknown. In this short review, we describe two distinct tilapia gene loci that encode Prl177 and Prl188 . From our in silico analyses of prl177 and prl188 promoter regions (approximately 1000 bp) and a transcriptome analysis of RPDs from fresh water (FW)- and seawater (SW)-acclimated tilapia, we propose a working model for how multiple transcription factors link osmoreceptive processes with adaptive patterns of prl177 and prl188 gene expression. We confirmed via RNA-sequencing and a quantitative polymerase chain reaction that multiple transcription factors emerging as predicted regulators of prl gene expression are expressed in the RPD of tilapia. In particular, gene transcripts encoding pou1f1, stat3, creb3l1, pbxip1a and stat1a were highly expressed; creb3l1, pbxip1a and stat1a were elevated in fish acclimated to SW vs FW. Combined, our in silico and transcriptome analyses set a path for resolving how adaptive patterns of Prl secretion are achieved via the integration of osmoreceptive processes with the control of prl gene transcription.

Ultra-Deep Sequencing Analysis on HIV Drug-Resistance-Associated Mutations Among HIV-Infected Individuals: First Report from the Philippines

A sharp increase in the number of people living with HIV has been documented in the Philippines. In response, the government has instituted antiretroviral therapy (ART) nationwide through HIV treatment hubs. However, no data presently exist on the status of ART drug-resistance-associated mutations (DRMs). In this study, we aim at analyzing DRM profiles in the Philippines and at providing comprehensive data on DRMs to guide treatment decisions and prevent viral failures. We conducted a cross-sectional study in 119 volunteers who tested positive for HIV from more than 8,000 participants screened for HIV across the nation through the 2013 Integrated HIV Behavioral and Serologic Surveillance (IHBSS) program. Amplicons were generated from plasma RNA by using primers designed to analyze diverse HIV-1 isolates targeting the reverse transcriptase region and sequenced on a 454 ultra-deep sequencing (UDS) platform to assess DRMs. DRMs were defined by using the Stanford HIV drug resistance database, and we found only 2 from 110 evaluable individuals with major HIV variants (>20% prevalence) that were highly resistant to the non-nucleoside reverse transcriptase inhibitor (NNRTI: efavirenz and nevirapine). However, a larger fraction of individuals harbored minority drug-resistant HIV variants (0.5%-20% prevalence) and they were highly resistant to NNRTI nevirapine (89/110), rilpivirine (5/110), and efavirenz (49/110). This study is the first report on the presence of HIV drug resistance in the Philippines and demonstrates the utility of UDS in assisting the detection of HIV minor variants. Monitoring for ART-DRMs will assist in improving HIV management strategies in curtailing the evolving epidemic in the Philippines.

Comparative genomics explains the evolutionary success of reef-forming corals

Transcriptome and genome data from twenty stony coral species and a selection of reference bilaterians were studied to elucidate coral evolutionary history. We identified genes that encode the proteins responsible for the precipitation and aggregation of the aragonite skeleton on which the organisms live, and revealed a network of environmental sensors that coordinate responses of the host animals to temperature, light, and pH. Furthermore, we describe a variety of stress-related pathways, including apoptotic pathways that allow the host animals to detoxify reactive oxygen and nitrogen species that are generated by their intracellular photosynthetic symbionts, and determine the fate of corals under environmental stress. Some of these genes arose through horizontal gene transfer and comprise at least 0.2% of the animal gene inventory. Our analysis elucidates the evolutionary strategies that have allowed symbiotic corals to adapt and thrive for hundreds of millions of years.

DOI:http://dx.doi.org/10.7554/eLife.13288.001

For millions of years, reef-building stony corals have created extensive habitats for numerous marine plants and animals in shallow tropical seas. Stony corals consist of many small, tentacled animals called polyps. These polyps secrete a mineral called aragonite to create the reef – an external ‘skeleton’ that supports and protects the corals.

Photosynthesizing algae live inside the cells of stony corals, and each species depends on the other to survive. The algae produce the coral’s main source of food, although they also produce some waste products that can harm the coral if they build up inside cells. If the oceans become warmer and more acidic, the coral are more likely to become stressed and expel the algae from their cells in a process known as coral bleaching. This makes the coral more likely to die or become diseased. Corals have survived previous periods of ocean warming, although it is not known how they evolved to do so.

The evolutionary history of an organism can be traced by studying its genome – its complete set of DNA – and the RNA molecules encoded by these genes. Bhattacharya et al. performed this analysis for twenty stony coral species, and compared the resulting genome and RNA sequences with the genomes of other related marine organisms, such as sea anemones and sponges. In particular, Bhattacharya et al. examined “ortholog” groups of genes, which are present in different species and evolved from a common ancestral gene. This analysis identified the genes in the corals that encode the proteins responsible for constructing the aragonite skeleton. The coral genome also encodes a network of environmental sensors that coordinate how the polyps respond to temperature, light and acidity.

Bhattacharya et al. also uncovered a variety of stress-related pathways, including those that detoxify the polyps of the damaging molecules generated by algae, and the pathways that enable the polyps to adapt to environmental stress. Many of these genes were recruited from other species in a process known as horizontal gene transfer.

The oceans are expected to become warmer and more acidic in the coming centuries. Provided that humans do not physically destroy the corals’ habitats, the evidence found by Bhattacharya et al. suggests that the genome of the corals contains the diversity that will allow them to adapt to these new conditions.

DOI:http://dx.doi.org/10.7554/eLife.13288.002

Preclinical development of HIvax: Human survivin highly immunogenic vaccines

Our previous work involved the development of a recombinant fowlpox virus encoding survivin (FP-surv) vaccine that was evaluated for efficacy in mesothelioma mouse models. Results showed that FP-surv vaccination generated significant immune responses, which led to delayed tumor growth and improved animal survival. We have extended those previous findings in the current study, which involves the pre-clinical development of an optimized version of FP-surv designed for human immunization (HIvax). Survivin-derived peptides for the most common haplotypes in the human population were identified and their immunogenicity confirmed in co-culture experiments using dendritic cells and T cells isolated from healthy donors. Peptides confirmed to induce CD8(+) and CD4(+) T cells activation in humans were then included in 2 transgenes optimized for presentation of processed peptides on MHC-I (HIvax1) and MHC-II (HIvax2). Fowlpox vectors expressing the HIvax transgenes were then generated and their efficacy was evaluated with subsequent co-culture experiments to measure interferon-γ and granzyme B secretion. In these experiments, both antigen specific CD4(+) and CD8(+) T cells were activated by HIvax vaccines with resultant cytotoxic activity against survivin-overexpressing mesothelioma cancer cells. These results provide a rationale for clinical testing of HIvax1 and HIvax2 vaccines in patients with survivin-expressing cancers.

Characterization of class II β chain major histocompatibility complex genes in a family of Hawaiian honeycreepers: 'amakihi (Hemignathus virens)

Hawaiian honeycreepers (Drepanidinae) have evolved in the absence of mosquitoes for over five million years. Through human activity, mosquitoes were introduced to the Hawaiian archipelago less than 200 years ago. Mosquito-vectored diseases such as avian malaria caused by Plasmodium relictum and Avipoxviruses have greatly impacted these vulnerable species. Susceptibility to these diseases is variable among and within species. Due to their function in adaptive immunity, the role of major histocompatibility complex genes (Mhc) in disease susceptibility is under investigation. In this study, we evaluate gene organization and levels of diversity of Mhc class II β chain genes (exon 2) in a captive-reared family of Hawaii 'amakihi (Hemignathus virens). A total of 233 sequences (173 bp) were obtained by PCR+1 amplification and cloning, and 5720 sequences were generated by Roche 454 pyrosequencing. We report a total of 17 alleles originating from a minimum of 14 distinct loci. We detected three linkage groups that appear to represent three distinct haplotypes. Phylogenetic analysis revealed one variable cluster resembling classical Mhc sequences (DAB) and one highly conserved, low variability cluster resembling non-classical Mhc sequences (DBB). High net evolutionary divergence values between DAB and DBB resemble that seen between chicken BLB system and YLB system genes. High amino acid identity among non-classical alleles from 12 species of passerines (DBB) and four species of Galliformes (YLB) was found, suggesting that these non-classical passerine sequences may be related to the Galliforme YLB sequences.

Identification and preliminary characterization of chemosensory perception-associated proteins in the melon fly Bactrocera cucurbitae using RNA-seq

An investigation into proteins involved in chemosensory perception in the melon fly, Bactrocera cucurbitae (Diptera: Tephritidae) is described here using a newly generated transcriptome dataset. The melon fly is a major agricultural pest, widely distributed in the Asia-Pacific region and some parts of Africa. For this study, a transcriptome dataset was generated using RNA extracted from 4-day-old adult specimens of the melon fly. The dataset was assembled and annotated via Gene Ontology (GO) analysis. Based on this and similarity searches to data from other species, a number of protein sequences putatively involved in chemosensory reception were identified and characterized in the melon fly. This included the highly conserved "Orco" along with a number of other less conserved odorant binding protein sequences. In addition, several sequences representing putative ionotropic and gustatory receptors were also identified. This study provides a foundation for future functional studies of chemosensory proteins in the melon fly and for making more detailed comparisons to other species. In the long term, this will ultimately help in the development of improved tools for pest management.

Demystifying computer science for molecular ecologists

In this age of data-driven science and high-throughput biology, computational thinking is becoming an increasingly important skill for tackling both new and long-standing biological questions. However, despite its obvious importance and conspicuous integration into many areas of biology, computer science is still viewed as an obscure field that has, thus far, permeated into only a few of the biology curricula across the nation. A national survey has shown that lack of computational literacy in environmental sciences is the norm rather than the exception [Valle & Berdanier (2012) Bulletin of the Ecological Society of America, 93, 373-389]. In this article, we seek to introduce a few important concepts in computer science with the aim of providing a context-specific introduction aimed at research biologists. Our goal was to help biologists understand some of the most important mainstream computational concepts to better appreciate bioinformatics methods and trade-offs that are not obvious to the uninitiated.

Chimeric piggyBac transposases for genomic targeting in human cells

Integrating vectors such as viruses and transposons insert transgenes semi-randomly and can potentially disrupt or deregulate genes. For these techniques to be of therapeutic value, a method for controlling the precise location of insertion is required. The piggyBac (PB) transposase is an efficient gene transfer vector active in a variety of cell types and proven to be amenable to modification. Here we present the design and validation of chimeric PB proteins fused to the Gal4 DNA binding domain with the ability to target transgenes to pre-determined sites. Upstream activating sequence (UAS) Gal4 recognition sites harbored on recipient plasmids were preferentially targeted by the chimeric Gal4-PB transposase in human cells. To analyze the ability of these PB fusion proteins to target chromosomal locations, UAS sites were randomly integrated throughout the genome using the Sleeping Beauty transposon. Both N- and C-terminal Gal4-PB fusion proteins but not native PB were capable of targeting transposition nearby these introduced sites. A genome-wide integration analysis revealed the ability of our fusion constructs to bias 24% of integrations near endogenous Gal4 recognition sequences. This work provides a powerful approach to enhance the properties of the PB system for applications such as genetic engineering and gene therapy.

The evolution of the tape measure protein: units, duplications and losses

BACKGROUND

A large family of viruses that infect bacteria, called phages, is characterized by long tails used to inject DNA into their victims' cells. The tape measure protein got its name because the length of the corresponding gene is proportional to the length of the phage's tail: a fact shown by actually copying or splicing out parts of DNA in exemplar species. A natural question is whether there exist units for these tape measures, and if different tape measures have different units and lengths. Such units would allow us to retrace the evolution of tape measure proteins using their duplication/loss history. The vast number of sequenced phages genomes allows us to attack this problem with a comparative genomics approach.

RESULTS

Here we describe a subset of phages whose tape measure proteins contain variable numbers of an 11 amino acids sequence repeat, aligned with sequence similarity, structural properties, and simple arithmetics. This subset provides a unique opportunity for the combinatorial study of phage evolution, without the added uncertainties of multiple alignments, which are trivial in this case, or of protein functions, that are well established. We give a heuristic that reconstructs the duplication history of these sequences, using divergent strains to discriminate between mutations that occurred before and after speciation, or lineage divergence. The heuristic is based on an efficient algorithm that gives an exhaustive enumeration of all possible parsimonious reconstructions of the duplication/speciation history of a single nucleotide. Finally, we present a method that allows, when possible, to discriminate between duplication and loss events.

CONCLUSIONS

Establishing the evolutionary history of viruses is difficult, in part due to extensive recombinations and gene transfers, and high mutation rates that often erase detectable similarity between homologous genes. In this paper, we introduce new tools to address this problem.

Mosaic graphs and comparative genomics in phage communities

Comparing the genomes of two closely related viruses often produces mosaics where nearly identical sequences alternate with sequences that are unique to each genome. When several closely related genomes are compared, the unique sequences are likely to be shared with third genomes, leading to virus mosaic communities. Here we present comparative analysis of sets of Staphylococcus aureus phages that share large identical sequences with up to three other genomes, and with different partners along their genomes. We introduce mosaic graphs to represent these complex recombination events, and use them to illustrate the breath and depth of sequence sharing: some genomes are almost completely made up of shared sequences, while genomes that share very large identical sequences can adopt alternate functional modules. Mosaic graphs also allow us to identify breakpoints that could eventually be used for the construction of recombination networks. These findings have several implications on phage metagenomics assembly, on the horizontal gene transfer paradigm, and more generally on the understanding of the composition and evolutionary dynamics of virus communities.

Divide and conquer: enriching environmental sequencing data

BACKGROUND

In environmental sequencing projects, a mix of DNA from a whole microbial community is fragmented and sequenced, with one of the possible goals being to reconstruct partial or complete genomes of members of the community. In communities with high diversity of species, a significant proportion of the sequences do not overlap any other fragment in the sample. This problem will arise not only in situations with a relatively even distribution of many species, but also when the community in a particular environment is routinely dominated by the same few species. In the former case, no genomes may be assembled at all, while in the latter case a few dominant species in an environment will always be sequenced at high coverage to the detriment of coverage of the greater number of sparse species.

METHODS AND RESULTS

Here we show that, with the same global sequencing effort, separating the species into two or more sub-communities prior to sequencing can yield a much higher proportion of sequences that can be assembled. We first use the Lander-Waterman model to show that, if the expected percentage of singleton sequences is higher than 25%, then, under the uniform distribution hypothesis, splitting the community is always a wise choice. We then construct simulated microbial communities to show that the results hold for highly non-uniform distributions. We also show that, for the distributions considered in the experiments, it is possible to estimate quite accurately the relative diversity of the two sub-communities.

CONCLUSION

Given the fact that several methods exist to split microbial communities based on physical properties such as size, density, surface biochemistry, or optical properties, we strongly suggest that groups involved in environmental sequencing, and expecting high diversity, consider splitting their communities in order to maximize the information content of their sequencing effort.

Wheat EST resources for functional genomics of abiotic stress

Background

Wheat is an excellent species to study freezing tolerance and other abiotic stresses. However, the sequence of the wheat genome has not been completely characterized due to its complexity and large size. To circumvent this obstacle and identify genes involved in cold acclimation and associated stresses, a large scale EST sequencing approach was undertaken by the Functional Genomics of Abiotic Stress (FGAS) project.

Results

We generated 73,521 quality-filtered ESTs from eleven cDNA libraries constructed from wheat plants exposed to various abiotic stresses and at different developmental stages. In addition, 196,041 ESTs for which tracefiles were available from the National Science Foundation wheat EST sequencing program and DuPont were also quality-filtered and used in the analysis. Clustering of the combined ESTs with d2_cluster and TGICL yielded a few large clusters containing several thousand ESTs that were refractory to routine clustering techniques. To resolve this problem, the sequence proximity and "bridges" were identified by an e-value distance graph to manually break clusters into smaller groups. Assembly of the resolved ESTs generated a 75,488 unique sequence set (31,580 contigs and 43,908 singletons/singlets). Digital expression analyses indicated that the FGAS dataset is enriched in stress-regulated genes compared to the other public datasets. Over 43% of the unique sequence set was annotated and classified into functional categories according to Gene Ontology.

Conclusion

We have annotated 29,556 different sequences, an almost 5-fold increase in annotated sequences compared to the available wheat public databases. Digital expression analysis combined with gene annotation helped in the identification of several pathways associated with abiotic stress. The genomic resources and knowledge developed by this project will contribute to a better understanding of the different mechanisms that govern stress tolerance in wheat and other cereals.