Comparison of 16S and whole genome dog microbiomes using machine learning.

BACKGROUND: Recent advances in sequencing technologies have driven studies identifying the microbiome as a key regulator of overall health and disease in the host. Both 16S amplicon and whole genome shotgun sequencing technologies are currently being used to investigate this relationship, however, the choice of sequencing technology often depends on the nature and experimental design of the study. In principle, the outputs rendered by analysis pipelines are heavily influenced by the data used as input; it is then important to consider that the genomic features produced by different sequencing technologies may emphasize different results. RESULTS: In this work, we use public 16S amplicon and whole genome shotgun sequencing (WGS) data from the same dogs to investigate the relationship between sequencing technology and the captured gut metagenomic landscape in dogs. In our analyses, we compare the taxonomic resolution at the species and phyla levels and benchmark 12 classification algorithms in their ability to accurately identify host phenotype using only taxonomic relative abundance information from 16S and WGS datasets with identical study designs. Our best performing model, a random forest trained by the WGS dataset, identified a species (Bacteroides coprocola) that predominantly contributes to the abundance of leuB, a gene involved in branched chain amino acid biosynthesis; a risk factor for glucose intolerance, insulin resistance, and type 2 diabetes. This trend was not conserved when we trained the model using 16S sequencing profiles from the same dogs. CONCLUSIONS: Our results indicate that WGS sequencing of dog microbiomes detects a greater taxonomic diversity than 16S sequencing of the same dogs at the species level and with respect to four gut-enriched phyla levels. This difference in detection does not significantly impact the performance metrics of machine learning algorithms after down-sampling. Although the important features extracted from our best performing model are not conserved between the two technologies, the important features extracted from either instance indicate the utility of machine learning algorithms in identifying biologically meaningful relationships between the host and microbiome community members. In conclusion, this work provides the first systematic machine learning comparison of dog 16S and WGS microbiomes derived from identical study designs.

An Assessment of the Stability of the Canine Oral Microbiota After Probiotic Administration in Healthy Dogs Over Time.

The administration of an oral probiotic has been demonstrated to impact oral microbial diversity in humans but has not been examined in canines. The objective of this study was to test the hypothesis that oral probiotic administration would impact the oral microbiota of canines compared to control. Working canines in training (n = 13) were assigned to Test or Control groups and acclimated to one of three commercially available study diets utilizing common protein sources (Purina Pro Plan Savor lamb, Purina Pro Plan Sport chicken, Purina Pro Plan Focus salmon) for a minimum of 30 days prior to initiation of the study. Following acclimation, dogs in the Test group began a daily regimen of oral probiotic (Fortiflora® Purina, St. Louis, MO) top-dressed on their midday feeding. Control dogs received their midday feeding with no probiotic. All dogs were sampled once weekly via oral pediatric swabs across the 7-week study. Next generation sequencing (Illumina, MiSeq) was utilized to develop microbial profiles specific to treatment, diet, and time. Bacterial composition was dominated by eight phyla (Proteobacteria 43.8%, Bacteroidetes 22.5%, Firmicutes 18.9%, Actinobacteria 6.1%, Fusobacteria 3.6%, Gracilibacteria 2.1%, SR1 Absconditabacteria 1.5%, and Saccharibacteria 1.3%) representing more than 99% of the relative abundance of the microbial composition. Probiotic administration failed to impact relative abundance at any taxonomic level (P > 0.05). Similarly, no effect on the oral microbiota was measured for diet (P > 0.05). Comparison using a Jaccard Index demonstrate a consistent microbial profile over the 7-week study with no impact evidenced by study week (P = 0.19). The data also revealed a profile of ubiquitous taxa that were present across all dogs and all samples regardless of breed, sex, diet, treatment or other factors. These genera include Actinomyces, Corynebacterium, Capnocytophaga, Flavobacterium, Gemella, Abiotrophia, Streptococcus, and Frederiksenia. These data demonstrate the stability of canine oral microbiota over time.

A Protein Antagonist of Activation-Induced Cytidine Deaminase Encoded by a Complex Mouse Retrovirus.

Complex human-pathogenic retroviruses cause high morbidity and mortality worldwide, but resist antiviral drugs and vaccine development due to evasion of the immune response. A complex retrovirus, mouse mammary tumor virus (MMTV), requires replication in B and T lymphocytes for mammary gland transmission and is antagonized by the innate immune restriction factor murine Apobec3 (mA3). To determine whether the regulatory/accessory protein Rem affects innate responses to MMTV, a splice-donor mutant (MMTV-SD) lacking Rem expression was injected into BALB/c mice. Mammary tumors induced by MMTV-SD had a lower proviral load, lower incidence, and longer latency than mammary tumors induced by wild-type MMTV (MMTV-WT). MMTV-SD proviruses had many G-to-A mutations on the proviral plus strand, but also C-to-T transitions within WRC motifs. Similarly, a lymphomagenic MMTV variant lacking Rem expression showed decreased proviral loads and increased WRC motif mutations relative to those in wild-type-virus-induced tumors, consistent with activation-induced cytidine deaminase (AID) mutagenesis in lymphoid cells. These mutations are typical of the Apobec family member AID, a B-cell-specific mutagenic protein involved in antibody variable region hypermutation. In contrast, mutations in WRC motifs and proviral loads were similar in MMTV-WT and MMTV-SD proviruses from tumors in AID-insufficient mice. AID was not packaged in MMTV virions. Rem coexpression in transfection experiments led to AID proteasomal degradation. Our data suggest that rem specifies a human-pathogenic immunodeficiency virus type 1 (HIV-1) Vif-like protein that inhibits AID and antagonizes innate immunity during MMTV replication in lymphocytes.IMPORTANCE Complex retroviruses, such as human-pathogenic immunodeficiency virus type 1 (HIV-1), cause many human deaths. These retroviruses produce lifelong infections through viral proteins that interfere with host immunity. The complex retrovirus mouse mammary tumor virus (MMTV) allows for studies of host-pathogen interactions not possible in humans. A mutation preventing expression of the MMTV Rem protein in two different MMTV strains decreased proviral loads in tumors and increased viral genome mutations typical of an evolutionarily ancient enzyme, AID. Although the presence of AID generally improves antibody-based immunity, it may contribute to human cancer progression. We observed that coexpression of MMTV Rem and AID led to AID destruction. Our results suggest that Rem is the first known protein inhibitor of AID and that further experiments could lead to new disease treatments.

Investigating Colonization of the Healthy Adult Gastrointestinal Tract by Fungi.

A wide diversity of fungi have been detected in the human gastrointestinal (GI) tract with the potential to provide or influence important functions. However, many of the fungi most commonly detected in stool samples are also present in food or the oral cavity. Therefore, to recognize which gut fungi are likely to have a sustained influence on human health, there is a need to separate transient members of the GI tract from true colonizers. To identify colonizing fungi, the eukaryotic rRNA operon's second internal transcribed spacer (ITS2) was sequenced from the stool, saliva, and food of healthy adults following consumption of different controlled diets. Unlike most bacterial 16S rRNA genes, the only fungal ITS2 operational taxonomic units (OTUs) detected in stool DNA across multiple diets were also present in saliva and/or food. Additional analyses, including culture-based approaches and sequencing of the 18S rRNA gene, ITS2 cDNA, and DNA extracted using alternative methods, failed to detect additional fungi. Two abundant fungi, Saccharomyces cerevisiae and Candida albicans, were examined further in healthy volunteers. Saccharomyces became undetectable in stool when a S. cerevisiae-free diet was consumed, and the levels of C. albicans in stool were dramatically reduced by more frequent cleaning of teeth. Extremely low fungal abundance, the inability of fungi to grow under conditions mimicking the distal gut, and evidence from analysis of other public datasets further support the hypothesis that fungi do not routinely colonize the GI tracts of healthy adults. IMPORTANCE We sought to identify the fungi that colonize healthy GI tracts and that have a sustained influence on the diverse functions of the gut microbiome. Instead, we found that all fungi in the stool of healthy volunteers could be explained by their presence in oral and dietary sources and that our results, together with those from other analyses, support the model that there is little or no gastrointestinal colonization by fungi. This may be due to Westernization, primate evolution, fungal ecology, and/or the strong defenses of a healthy immune system. Importantly, fungal colonization of the GI tract may often be indicative of disease. As fungi can cause serious infections in immunocompromised individuals and are found at increased abundance in multiple disorders of the GI tract, understanding normal fungal colonization is essential for proper treatment and prevention of fungal pathogenesis.

The gut mycobiome of the Human Microbiome Project healthy cohort.

BACKGROUND: Most studies describing the human gut microbiome in healthy and diseased states have emphasized the bacterial component, but the fungal microbiome (i.e., the mycobiome) is beginning to gain recognition as a fundamental part of our microbiome. To date, human gut mycobiome studies have primarily been disease centric or in small cohorts of healthy individuals. To contribute to existing knowledge of the human mycobiome, we investigated the gut mycobiome of the Human Microbiome Project (HMP) cohort by sequencing the Internal Transcribed Spacer 2 (ITS2) region as well as the 18S rRNA gene. RESULTS: Three hundred seventeen HMP stool samples were analyzed by ITS2 sequencing. Fecal fungal diversity was significantly lower in comparison to bacterial diversity. Yeast dominated the samples, comprising eight of the top 15 most abundant genera. Specifically, fungal communities were characterized by a high prevalence of Saccharomyces, Malassezia, and Candida, with S. cerevisiae, M. restricta, and C. albicans operational taxonomic units (OTUs) present in 96.8, 88.3, and 80.8% of samples, respectively. There was a high degree of inter- and intra-volunteer variability in fungal communities. However, S. cerevisiae, M. restricta, and C. albicans OTUs were found in 92.2, 78.3, and 63.6% of volunteers, respectively, in all samples donated over an approximately 1-year period. Metagenomic and 18S rRNA gene sequencing data agreed with ITS2 results; however, ITS2 sequencing provided greater resolution of the relatively low abundance mycobiome constituents. CONCLUSIONS: Compared to bacterial communities, the human gut mycobiome is low in diversity and dominated by yeast including Saccharomyces, Malassezia, and Candida. Both inter- and intra-volunteer variability in the HMP cohort were high, revealing that unlike bacterial communities, an individual's mycobiome is no more similar to itself over time than to another person's. Nonetheless, several fungal species persisted across a majority of samples, evidence that a core gut mycobiome may exist. ITS2 sequencing data provided greater resolution of the mycobiome membership compared to metagenomic and 18S rRNA gene sequencing data, suggesting that it is a more sensitive method for studying the mycobiome of stool samples.

How Close Are We to Achieving Energy and Nutrient Goals for Very Low Birth Weight Infants in the First Week?

BACKGROUND: Emerging evidence suggests intakes of protein and energy as early as the first week of life in preterm very low birth weight (VLBW) infants are associated with improved neurodevelopment. In response, many neonatal intensive care units (NICUs) have launched new, more aggressive early feeding guidelines. The aim of this study was to evaluate enteral and parenteral energy and macronutrient intakes during the first postnatal week in VLBW infants admitted to NICUs that have introduced more aggressive early feeding guidelines. MATERIALS AND METHODS: Estimated energy and macronutrient intakes were prospectively collected from VLBW infants fed exclusively mother's own milk and/or parenteral nutrition and compared with expert recommendations. Days to reach full enteral feeds (150 mL/kg/d) and discharge anthropometrics were examined. RESULTS: By days 6 and 7, median protein and lipid intakes, respectively, reached recommended values (3.5 and 3.0 g/kg/d). However, by day 8, many infants remained below recommended intakes for protein (34%), lipid (34%), carbohydrate (68%), and energy (71%). Late-onset sepsis was associated with a decreased likelihood of reaching full enteral feeds on any given day (hazard ratio, 0.2; 95% confidence interval, 0.1-0.5; P ≤ .0009). There was no significant relationship between week 1 nutrient intakes and anthropometrics at discharge. CONCLUSION: Despite the introduction of more aggressive early feeding guidelines and improved energy and nutrient intakes compared with literature values, many VLBW infants remain below recommended nutrition goals in the first week.

Retroviral vectors elevate coexpressed protein levels in trans through cap-dependent translation.

Retroviruses cause immunodeficiency and cancer but also are used as vectors for the expression of heterologous genes. Nevertheless, optimal translation of introduced genes often is not achieved. Here we show that transfection into mammalian cells of lentiviral or gammaretroviral vectors, including those with specific shRNAs, increased expression of a cotransfected gene relative to standard plasmid vectors. Levels of most endogenous cellular proteins were unchanged. Transfer of lentiviral vector sequences into a standard plasmid conferred the ability to give increased expression of cotransfected genes (superinduction). Superinduction by the retroviral vector was not dependent on the cell type or species, the type of reporter gene, or the method of transfection. No differences were detected in the IFN, unfolded protein, or stress responses in the presence of retroviral vectors. RT-PCRs revealed that RNA levels of cotransfected genes were unchanged during superinduction, yet Western blotting, pulse labeling, and the use of bicistronic vectors showed increased cap-dependent translation of cointroduced genes. Expression of the mammalian target of rapamycin (mTOR) kinase target 4E-BP1, but not the mTOR inhibitor Torin 1, preferentially inhibited superinduction relative to basal protein expression. Furthermore, transcription of lentiviral vector sequences from a doxycycline-inducible promoter eliminated superinduction, consistent with a DNA-triggered event. Thus, retroviral DNA increased translation of cointroduced genes in trans by an mTOR-independent signaling mechanism. Our experiments have broad applications for the design of retroviral vectors for transfections, DNA vaccines, and gene therapy.

Fungal morphogenetic pathways are required for the hallmark inflammatory response during Candida albicans vaginitis.

Vulvovaginal candidiasis, caused primarily by Candida albicans, presents significant health issues for women of childbearing age. As a polymorphic fungus, the ability of C. albicans to switch between yeast and hyphal morphologies is considered its central virulence attribute. Armed with new criteria for defining vaginitis immunopathology, the purpose of this study was to determine whether the yeast-to-hypha transition is required for the hallmark inflammatory responses previously characterized during murine vaginitis. Kinetic analyses of vaginal infection with C. albicans in C57BL/6 mice demonstrated that fungal burdens remained constant throughout the observation period, while polymorphonuclear leukocyte (PMN), S100A8, and interleukin-1ß levels obtained from vaginal lavage fluid increased by day 3 onward. Lactate dehydrogenase activity was also positively correlated with increased effectors of innate immunity. Additionally, immunodepletion of neutrophils in infected mice confirmed a nonprotective role for PMNs during vaginitis. Determination of the importance of fungal morphogenesis during vaginitis was addressed with a two-pronged approach. Intravaginal inoculation of mice with C. albicans strains deleted for key transcriptional regulators (bcr1Δ/Δ, efg1Δ/Δ, cph1Δ/Δ, and efg1Δ/Δ cph1Δ/Δ) controlling the yeast-to-hypha switch revealed a crucial role for morphogenetic signaling through the Efg1 and, to a lesser extent, the Bcr1 pathways in contributing to vaginitis immunopathology. Furthermore, overexpression of transcription factors NRG1 and UME6, to maintain yeast and hyphal morphologies, respectively, confirmed the importance of morphogenesis in generating innate immune responses in vivo. These results highlight the yeast-to-hypha switch and the associated morphogenetic response as important virulence components for the immunopathogenesis of Candida vaginitis, with implications for transition from benign colonization to symptomatic infection.

Requirements for mouse mammary tumor virus Rem signal peptide processing and function.

Mouse mammary tumor virus (MMTV) encodes a Rev-like protein, Rem, which is involved in the nuclear export and expression of viral RNA. Previous data have shown that all Rev-like functions are localized to the 98-amino-acid signal peptide (SP) at the N terminus of MMTV Rem or envelope proteins. MMTV-SP uses endoplasmic reticulum-associated degradation (ERAD) for protein trafficking. Rem cleavage by signal peptidase in the ER is necessary for MMTV-SP function in a reporter assay, but many requirements for trafficking are not known. To allow detection and localization of both MMTV-SP and the C-terminal cleavage product, we prepared plasmids expressing green fluorescent protein (GFP) tags. N-terminal Rem tagging led to protein accumulation relative to untagged Rem and allowed signal peptidase cleavage but reduced its specific activity. C-terminal tagging also led to Rem accumulation yet dramatically reduced cleavage, GFP fluorescence, and activity relative to N-terminally tagged Rem (GFPRem). Substitutions of an invariant leucine at position 71 between the known RNA-binding and nuclear export sequences interfered with GFPRem accumulation and activity but not cleavage. Similarly, deletion of 100 or 150 C-terminal amino acids from GFPRem dramatically reduced both Rem and MMTV-SP levels and function. Removal of the entire C terminus (203 amino acids) restored both protein levels and activity of MMTV-SP. Only C-terminal GFP tagging, and not other modifications, appeared to trap Rem in the ER membrane. Thus, Rem conformation in both the ER lumen and cytoplasm determines cleavage, retrotranslocation, and MMTV-SP function. These mutants further characterize intermediates in Rem trafficking and have implications for all proteins affected by ERAD.

Pattern of growth of very low birth weight preterm infants, assessed using the WHO Growth Standards, is associated with neurodevelopment.

Several Canadian professional organizations recently recommended that the growth of preterm infants be monitored using the World Health Organization Growth Standards (WHO-GS) after hospital discharge. The WHO-GS are a prescriptive set of growth charts that describe how term infants should grow under ideal environmental conditions. Whether preterm infants following this pattern of growth have better outcomes than infants that do not has yet to be evaluated. Our aim was to determine whether the pattern of growth of very low birth weight (VLBW) infants during the first 2 years, assessed using the WHO-GS or the traditional Centers for Disease Control and Prevention reference growth charts (CDC-RGC), is associated with neurodevelopment. Pattern of weight, length, and head circumference gain of appropriate-for-gestation VLBW preterm infants (n = 289) from birth to 18-24 months corrected age was classified, using the WHO-GS and CDC-RGC, as sustained (change in Z-score ≤1 SD), decelerated (decline >1 SD), or accelerated (incline >1 SD). Development was assessed using the Bayley Scales of Infant and Toddler Development (BSID)-III at 18-24 months corrected age. Using the WHO-GS, children with a decelerated pattern of weight gain had lower cognitive (10 points), language (6 points), and motor (4 points) scores than infants with sustained weight gain (p < 0.05), even after adjustment for morbidities. No association was found using the CDC-RGC. In conclusion, a decelerated pattern of weight gain, determined with the WHO-GS, but not the CDC-GRC, is associated with poorer neurodevelopment scores on the BSID-III than a pattern of sustained growth.

Role of ascorbate in lung cellular toxicity mediated by light-exposed parenteral nutrition solution.

Neonatal lung injury has been induced experimentally by infusion of multivitamin-containing light-exposed parenteral nutrition (PN) solutions. The objective was to explore the role of ascorbate in toxic effects of light-exposed PN on primary cultured foetal rat lung epithelial cells. Hydroperoxides were measured in 3% amino acid solutions at baseline, immediately after addition of either multivitamins or ascorbate alone (400 µg/mL) and again after a 24-h period of exposure to (or protection from) ambient light. Cellular toxicity was assessed by [C(14)]adenine release. Multivitamins or ascorbate alone increased hydroperoxides in PN, which was attenuated by light protection. Light-exposed PN containing multivitamins was more toxic to cells than baseline or light-protected PN. Exposure to ascorbate at concentrations both lower (< 5 µg/mL) and higher (> 1000 µg/mL) than normally contained in PN-induced oxidant-mediated cell death, as indicated by protective effects of hydroperoxide and hydroxyl radical scavengers. This study concludes that ascorbate generates toxic amounts of peroxide in PN solutions. The types and physiological importance of hydroperoxides induced by pro-oxidant effects of ascorbate require further evaluation in vivo.

Bar coding from breast to baby: a comprehensive breast milk management system for the NICU.

Breast milk errors have received increasing attention in the literature in terms of the potential infectious risk posed to the recipient baby and also the stress that results for both the donor and recipient families. Beginning in the mid-1990s, one Level III NICU began making changes in how feedings were prepared and distributed in an attempt to reduce breast milk errors. Despite these changes, breast milk errors continued to occur, and, in 2005, this NICU introduced a bar coding system to further reduce the risk of administering breast milk to the wrong infant. Breast milk errors have subsequently been substantially reduced.

Field testing of the 2006 World Health Organization growth charts from birth to 2 years: assessment of hospital undernutrition and overnutrition rates and the usefulness of BMI.

BACKGROUND: The World Health Organization (WHO) recently released a growth standard, a first attempt at describing how children should grow in an ideal environment. These charts introduce body mass index (BMI)-for-age percentiles for children younger than 2 years. Adopting the WHO standard may affect the number of children screened to require follow-up; hence, field testing needs to be completed in a tertiary care center where the incidence of suboptimal nutrition is high. The objectives of this study were to quantify differences between the new WHO and 2000 Centers for Disease Control and Prevention (CDC) growth charts for children younger than 2 years. The interchangeability of the WHO weight-for-length and WHO BMI percentiles was also assessed. METHODS: Percentile scores were computed for children younger than 2 years (n = 547) admitted to a pediatric tertiary health care center in Toronto, Canada. RESULTS: The WHO standard identified more children younger than 2 years as at risk of overweight/obesity compared with the CDC reference (21.0% vs 16.6%, >or=85th weight-for-length percentile) and fewer children as wasted (18.6% vs 23.0%, <5th weight-for-length percentile). The WHO BMI-for-age and WHO weight-for-length percentiles were highly correlated (r2 = 0.83) but not interchangeable. For approximately 9% of all children, and approximately 16% of those aged 25 percentile points. CONCLUSIONS: These data describe for the first time the magnitude of differences in the number of children screened as undernourished (4.4% decrease) or overnourished (4.4% increase) with adoption of the WHO standard in a tertiary care setting. Furthermore, the WHO's BMI-for-age and weight-for-length percentiles for children younger than 2 years are correlated but are not interchangeable.

Growth assessment in infants and toddlers using three different reference charts.

OBJECTIVE: To determine if the proportion of children < or =24 months old in a tertiary care facility defined as at risk of undernutrition or overnutrition differs according to different references used for assessment: the Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS) or Tanner-Whitehouse (Tanner) growth charts for weight-for-age and length-for-age. METHODS: Lengths and weights were measured on infants (207 female, 341 male) aged < or =24 months admitted from or attending clinics in the General Pediatric or Respiratory Medicine Programs at The Hospital for Sick Children, Toronto. Weight-for-age and length-for-age percentiles and percent ideal body weight were electronically computed. RESULTS: The proportion of all children whose weight-for-age was <3rd percentile (at risk of undernutrition) was greatest using the CDC growth charts (22.5%) compared with the NCHS (15.9%) or Tanner (19.2%) growth charts. Likewise, the proportion of all infants/toddlers with percent ideal body weight <90 (at risk of undernutrition) was greatest using the CDC (32.3%) compared with the NCHS (22.1%) or Tanner (25.9%) growth charts. In contrast, the percentage of children whose percent ideal body weight was > or =110% (at risk of overnutrition) was least using the CDC (18.1%) compared with the NCHS (26.1%) or Tanner (22.4%) growth charts. CONCLUSION: More children aged < or =24 months will be defined as at risk of undernutrition and fewer at risk of overnutrition when using weight-for-age or percent ideal body weight and the CDC growth charts compared with the NCHS or Tanner growth charts. As a result, requests for a more detailed nutritional assessment for undernutrition will likely follow implementation of the CDC growth charts in a tertiary care setting. As the CDC, NCHS and Tanner growth charts are growth "references" rather than "standards," other than for screening purposes, they should not be used in isolation when assessing growth and nutritional status.