Machine learning classification by fitting amplicon sequences to existing OTUs.

The ability to use 16S rRNA gene sequence data to train machine learning classification models offers the opportunity to diagnose patients based on the composition of their microbiome. In some applications, the taxonomic resolution that provides the best models may require the use of de novo operational taxonomic units (OTUs) whose composition changes when new data are added. We previously developed a new reference-based approach, OptiFit, that fits new sequence data to existing de novo OTUs without changing the composition of the original OTUs. While OptiFit produces OTUs that are as high quality as de novo OTUs, it is unclear whether this method for fitting new sequence data into existing OTUs will impact the performance of classification models relative to models trained and tested only using de novo OTUs. We used OptiFit to cluster sequences into existing OTUs and evaluated model performance in classifying a dataset containing samples from patients with and without colonic screen relevant neoplasia (SRN). We compared the performance of this model to standard methods including de novo and database-reference-based clustering. We found that using OptiFit performed as well or better in classifying SRNs. OptiFit can streamline the process of classifying new samples by avoiding the need to retrain models using reclustered sequences. IMPORTANCE There is great potential for using microbiome data to aid in diagnosis. A challenge with de novo operational taxonomic unit (OTU)-based classification models is that 16S rRNA gene sequences are often assigned to OTUs based on similarity to other sequences in the dataset. If data are generated from new patients, the old and new sequences must be reclustered to OTUs and the classification model retrained. Yet there is a desire to have a single, validated model that can be widely deployed. To overcome this obstacle, we applied the OptiFit clustering algorithm to fit new sequence data to existing OTUs allowing for reuse of the model. A random forest model implemented using OptiFit performed as well as the traditional reassign and retrain approach. This result shows that it is possible to train and apply machine learning models based on OTU relative abundance data that do not require retraining or the use of a reference database.

Infection-Induced Changes Within the Endocytic Recycling Compartment Suggest a Roadmap of Human Cytomegalovirus Egress.

Human cytomegalovirus (HCMV) is an important pathogen in developing fetuses, neonates, and individuals with compromised immune systems. Gaps in our understanding of the mechanisms required for virion assembly stand in the way of development of antivirals targeting late stages of viral replication. During infection, HCMV causes a dramatic reorganization of the host endosecretory system, leading to the formation of the cytoplasmic virion assembly complex (cVAC), the site of virion assembly. As part of cVAC biogenesis, the composition and behavior of endosecretory organelles change. To gain more comprehensive understanding of the impact HCMV infection has on components of the cellular endocytic recycling compartment (ERC), we used previously published transcriptional and proteomic datasets to predict changes in the directionality of ERC trafficking. We identified infection-associated changes in gene expression that suggest shifts in the balance between endocytic and exocytic recycling pathways, leading to formation of a secretory trap within the cVAC. Conversely, there was a corresponding shift favoring outbound secretory vesicle trafficking, indicating a potential role in virion egress. These observations are consistent with previous studies describing sequestration of signaling molecules, such as IL-6, and the synaptic vesicle-like properties of mature HCMV virions. Our analysis enabled development of a refined model incorporating old and new information related to the behavior of the ERC during HCMV replication. While limited by the paucity of integrated systems-level data, the model provides an informed basis for development of experimentally testable hypotheses related to mechanisms involved in HCMV virion maturation and egress. Information from such experiments will provide a robust roadmap for rational development of novel antivirals for HCMV and related viruses.

Betaherpesvirus Virion Assembly and Egress.

Virions are the vehicle for cell-to-cell and host-to-host transmission of viruses. Virions need to be assembled reliably and efficiently, be released from infected cells, survive in the extracellular environment during transmission, recognize and then trigger entry of appropriate target cells, and disassemble in an orderly manner during initiation of a new infection. The betaherpesvirus subfamily includes four human herpesviruses (human cytomegalovirus and human herpesviruses 6A, 6B, and 7), as well as viruses that are the basis of important animal models of infection and immunity. Similar to other herpesviruses, betaherpesvirus virions consist of four main parts (in order from the inside): the genome, capsid, tegument, and envelope. Betaherpesvirus genomes are dsDNA and range in length from ~145 to 240 kb. Virion capsids (or nucleocapsids) are geometrically well-defined vessels that contain one copy of the dsDNA viral genome. The tegument is a collection of several thousand protein and RNA molecules packed into the space between the envelope and the capsid for delivery and immediate activity upon cellular entry at the initiation of an infection. Betaherpesvirus envelopes consist of lipid bilayers studded with virus-encoded glycoproteins; they protect the virion during transmission and mediate virion entry during initiation of new infections. Here, we summarize the mechanisms of betaherpesvirus virion assembly, including how infection modifies, reprograms, hijacks, and otherwise manipulates cellular processes and pathways to produce virion components, assemble the parts into infectious virions, and then transport the nascent virions to the extracellular environment for transmission.

Generation of a novel human cytomegalovirus bacterial artificial chromosome tailored for transduction of exogenous sequences.

The study of herpesviruses, including human cytomegalovirus (HCMV), is complicated by viral genome complexity and inefficient methods for genetic manipulation in tissue culture. Reverse genetics of herpesviruses has been facilitated by propagating their genomes in E. coli as bacterial artificial chromosomes (BACs), which enables complex and precise genetic manipulation using bacterial recombinational systems. Internal capsid volume imposes a strict limit on the length of genome that can be packaged efficiently. This necessitates deletion of presumably nonessential segments of the viral genome to allow for incorporation of the E. coli mini-F plasmid propagation sequence. To avoid deleting viral genes, several BACs utilize a Cre/LoxP system to self-excise the mini-F sequence upon reconstitution of virus in tissue culture. Here, we describe the adaptation of Cre/LoxP to modify the mini-F sequence of the HCMV TB40/E BAC, thus generating a new self-excisable BAC, TB40/E/Cre. After excision of the E. coli propagation sequence, a 2.7 kbp genome length deficit is created due to a preexisting deletion within the US2-US6 coding region. We exploited this deficit and an FKBP12 protein destabilization domain (ddFKBP) to create a novel gene transduction system for studying exogenous proteins during HCMV infection. Using TB40/E/Cre, we: i) found genome length-associated differences in growth and ii) demonstrated its utility as a system capable of efficient transduction of exogenous proteins and regulation of their accumulation over periods as short as 2h. TB40/E/Cre is a powerful tool of broad applicability that can be adapted to study HCMV replication and cell biology in a variety of contexts.