Our bodies are home to millions and millions of microbes (bacteria, fungi and viruses), that live in harmony with us without producing any negative (disease producing) effects. Research is beginning to show that these microbes interact with us to help with our immune system, digestive tract and brain development among many other effects. This community of microbes, known collectively as our microbiome, may commence colonisation while we are developing in the womb and becomes quickly established after we are born. Much remains unknown about how preterm birth affects the development of our microbiome.

The goal of this longitudinal observational study is to gather more information of how and from where we get those first few microbes, the pattern in which our microbiome develops, and how intensive care for a preterm baby affects this.

The main questions it aims to answer are:

• How is the gut microbiome of a very premature infant affected by clinical management practices (e.g. antibiotics, probiotics, feeding) and how does it progress subsequently.
• How do probiotics colonise the preterm gut, and how do they persist once supplementation is discontinued.

Samples will be collected from mothers and their infants during the NICU admission including:

• A rectal swab
• Meconium and stool
• Urine
• Blood
• Expressed breastmilk
• Maternal stool
• Maternal oral swab
• Maternal vaginal or skin swab (depending on mode of delivery)

Samples will be analysed using next generation sequencing techniques to, for example, evaluate microbial composition of the samples or determine functional microbiome-host interactions.

Work done previously has identified several factors associated with alterations in the neonatal gut microbiome. These include gestational age, mode of delivery, exposure to antibiotics and feeding choice. Mothers of very low birth weight infants are typically encouraged to express breastmilk for their baby. This is associated with lower rates of sepsis and of necrotizing enterocolitis. Some of these effects may be related to the bacteria received in breastmilk from their mother. Correlations between changes in the neonatal gut microbiome and the microbiota of the milk they are being fed are limited. Moreover, the effect of freezing (i.e. standard breastmilk storage practice) on breastmilk microbiota has not been well characterized. Furthermore, in the absence of Mother's Own Milk (MOM) it is clinically appropriate to offer Donor Human Milk, where the additional process of pasteurization may also affect the bacterial strains present therein and lead to an altered pattern of colonization of the neonatal gut.

This is a prospective observational study which will evaluate the establishment and evolution of the intestinal microbiome of very preterm infants using next generation sequencing techniques. Data will be analyzed using bioinformatics to assess vertical transmission from mother to infant in utero, establishment and evolution of the preterm infant microbiome and evaluate how Neonatal Intensive Care Unit (NICU) management strategies alter the microbiome composition and recovery after interventions. This work will provide detailed information which may provide the basis for development of individualized probiotics for very preterm infants and may help develop microbiome protective management strategies for the NICU.

Recruitment and consent process Pregnant women will be approached if there is a possibility of delivery prior to 32 weeks gestation. A participant information leaflet will be provided which will detail the purpose of the study and what would be expected from participants. The study will also be explained by a researcher. Questions will be answered and sufficient time for deliberation will be provided. If they wish to participate with their infant, the putatively eligible participants must sign the informed consent form (ICF). They will receive a copy of the signed consent form. Deferred consent will be allowed but any infant samples collected will be stored on site and will not be processed prior to consent being obtained. If consent is declined any samples collected will be discarded immediately.

Inclusion/exclusion criteria will be assessed at all timepoints from screening to informed consent to delivery of the infant. Participants may be withdrawn from the study if exclusion criteria are met postnatally. For inclusion/exclusion criteria see Sections 4.1/4.2. Multiple births will be enrolled separately. Samples will be collected from each infant but only one set of maternal samples would be collected. In the event that one twin is delivered vaginally and the other by caesarean section both skin and vaginal swabs will be collected from the mother if possible.

Information regarding demographics, delivery, auxology, feeding, antibiotics and comorbidities will be collected from the participants. In addition, a Bayley scale of infant and toddler development will be carried out at 2 years of age to assess neurodevelopment.

Sample Collection

Infant samples

Meconium and stool samples Stool sampling will be carried out at the cotside. Rectal swabs, performed by the admitting clinical nurse, are taken routinely on admission to the NICU. An additional swab will be taken for microbiome analysis at this time. During the neonatal admission, a sample of the first meconium passed will be collected from the nappy and thereafter stool will be sampled from the infant's nappy bi-weekly, if possible, during routine cares.

Infant urine During the inpatient stay urine will be collected weekly by the neonatal nurse by placing a sterile pad/swabs in the infant's nappy which will then be removed at the next occurrence of routine cares.

Infant blood Blood samples will be collected only at the time of venepuncture for a clinically indicated reason. Venepuncture and sample collection will be carried out by the clinical team. No more than 1% of an infant's estimated total blood volume (eTBV) (80mls/kg) will be collected at any one timepoint for research purposes (eTBV=80mls/kg). Timepoints for blood sample collection will be within the first week of life (i.e. prior to establishment of feeding), between 2 and 4 weeks of life (i.e. once enteral feeding has been fully established), and prior to discharge. Venepuncture for the sole purpose of collecting a research sample will not be allowed.

Maternal Samples

Oral swabs Oral samples will be collected by the research nurse/investigator as soon as possible after birth from the mother using a sterile swab.

Vaginal swab (Vaginal deliveries) A vaginal swab will be collected from the mother by the attending midwife or obstetrician in the delivery suite, or by the mother herself in the postnatal ward.

Skin swab (Caesarean Section Deliveries) A skin swab will be obtained from the mother by the research nurse/investigator post-delivery. It will be collected from the internal aspect of wrist under the participant's hospital wrist band.

Maternal stool samples The participants will be provided with kits and instructions to collect stool samples. Samples will be frozen as soon as provided and collected by the research team.

Breast milk Breast milk samples of MOM will be collected from mothers throughout the study. Mothers will be asked to provide a 5-15 ml breastmilk sample (if possible) every two weeks while their infant is in NICU if she has an adequate supply.

Sample Analysis

Analysis for infant stool High-throughput screening of heterologously expressed metagenomic DNA will be examined as a means of looking at the functionality e.g. overall ability to metabolize certain carbohydrate sources (i.e. breast milk). DNA will be extracted from the feces of all infants followed by generation of PCR amplicons using universal 16s RNA primers, purification and quantification of PCR products. Illumina Miseq sequencing will then be carried out on 16S rRNA genes fragments per individual. Due to the vast amount of data that is generated from this technology, subsequent bioinformatic and statistical analysis is essential.

Analysis for infant urine Untargeted LC-MS assays will be performed with a hybrid linear ion trap Fourier Transform (LTQ FT) Orbitrap mass spectrometer. The XCMS Online portal (https://xcmsonline.scripps.edu/) will be used for data processing (alignment, peak picking, zero peak re-integrations, features grouping and assessment of quality control samples). To annotate compounds, a selection strategy will be used based on the most abundant and the most statistically significant features. To aid the identification of metabolites, an additional clustering analysis will be performed using Spearman correlation analysis.

Analysis for infant blood Blood samples will be examined in relation to functional metabolomics, including, but not limited to, analysis using shotgun metabolomics and comparison to the stool microbiome high throughput datasets at the same time points looking for relationships to known microbial signaling groups.

Analysis of maternal samples The peri-partum oral, vaginal and skin microbial composition, in conjunction with general clinical and demographic data, will be analyzed. Comparison will be made between maternal microbiota profiles with previous studies and data will be examined with the early infant gut microbiome data to assess the vertical and horizontal transfer of microorganisms in maternal-infant transmission. DNA will be extracted from samples, followed by generation of PCR amplicons using universal 16s RNA primers, purification and quantification of PCR products. Illumina Miseq sequencing will then be carried out on 16S rRNA genes fragments per individual. Due to the vast amount of data that is generated from this technology, subsequent bioinformatic and statistical analysis is essential.

Stool microbiota composition will be analyzed for their microbial content by the investigators using state-of-the art methods, possibly including but not limited to sequencing, and quantitative PCR.

The investigators will analyze maternal fecal microbial composition, and the fecal microbial metagenome, in conjunction with general clinical and demographic data. Fecal microbial compositional sequences will reveal which bacterial types are present, and metagenomic analyses will indicate the metabolic potential of the microbial load. We propose to use high-throughput screening of heterologously expressed metagenomic DNA as a means of looking at the functionality of different populations e.g. overall ability to metabolize certain carbohydrate sources, transformation of toxins etc. DNA will be extracted from the maternal feces as outlined for the infant stool.

Analysis for breast milk The investigators will analyze milk microbial composition, and microbial metagenome, in conjunction with general clinical and demographic data. Bacterial profiles will be compared against previous studies reported on the maternal gut microbiome and correlate the data with the infant stool microbiome data to assess to horizontal transfer of microorganisms in maternal-infant transmission. High-throughput screening of heterologously expressed metagenomic DNA will be used as a means of looking at the functionality e.g. overall ability to metabolize certain carbohydrate sources (e.g. human milk oligosaccharides). DNA will be extracted from the breast milk of all subjects, followed by generation of PCR amplicons using universal 16s RNA primers, purification and quantification of PCR products. Illumina Miseq sequencing will then be carried out on 16S rRNA genes fragments per individual. Bacteriophage are important regulators of microbiome composition and play a key role in maintaining the appropriate milk ecosystem equilibrium between predator and prey. Bacteriophage will be isolated from human milk for the purposes of sequence independent metagenomic sequencing. Those bacteriophage associated with immunoglobulins will be similarly isolated and sequenced. Due to the vast amount of data that is generated from this technology, subsequent bioinformatic and statistical analysis is essential. Bioinformatic analysis will be the same as mentioned previously for the other biological samples.

All analytical methods will be updated as appropriate in accordance with state-of-the-art methodology.

Sample Storage All maternal samples will be collected immediately by the researcher and frozen on site at -20°C until they can be transported to the off-site laboratory at APC Microbiome Ireland for processing as soon as possible and freezing for storage.

All infant samples obtained in the NICU will be stored in the research freezer until they can be transported to the off-site laboratory at APC Microbiome Ireland for processing as soon as possible and freezing for storage.

Bioinformatic analysis Bioinformatic analysis will involve the following, 300 bp paired-end reads will be assembled using FLASH and the QIIME suite of tools will be used for further processing of paired-end reads, including quality filtering and removal of mismatched barcodes and sequences below length thresholds. OTU sequences will be aligned using PyNAST and the SILVA SSURef database will be used to determine taxonomy. Alpha and beta diversities will be calculated using QIIME. Finally, principal coordinate analysis (PCoA) plots will be used to visualize differences in beta diversity based on UniFrac distance matrices.

Primary Endpoint The primary study endpoint will be the 24-month assessment of the last participant enrolled. The study aims to enrol 80 - 100 infants born before 32 weeks gestation over an 18-month period. Analyses will be performed from the first samples after birth to detect existing, early colonizers of the infant gut (likely through vertical transmission) and determine patterns of microbiome establishment in these preterm infants.

Regulatory & Administrative Procedures The study will be conducted in accordance with the version Fortaleza, Brazil, October 2013 of the Declaration of Helsinki 1964. The Protocol and the Participant Information Leaflet/ICF will be approved by the Clinical Research Ethics Committee of the Cork Teaching Hospitals.

Participant information/Informed consent The informed consent form used in this study, and any changes made during the study, have been prospectively approved by the Clinical Research Ethics Committee of the Cork Teaching Hospitals before use. Each informed consent form will include all relevant elements currently required by the ICH E6 Guideline for Good Clinical Practice (GCP) and Data Protection Act of 1988 and 2003, and GDPR from 25 May 2018.

Participants may withdraw from participation in the study at any time without detriment. Additionally, the investigator may withdraw participants from the study if it is in the best interest of the participant. The reason for participant withdrawals from the study will be documented in the participants case notes.

GCP and Record Retention The study will be managed and conducted according to the latest International Conference on Harmonisation (ICH) guidelines for Good Clinical Practice (GCP). The investigator will perform or directly supervise the performance of all the services described herein, or incidental to those described herein, in accordance to the highest standards of medical and clinical research practice. The use and storage of data collected during this study will comply with the General Data Protection Regulation (GDPR) (Regulation (EU) 2016/679 of the European Parliament and of the Council) and the Data Protection Act 2018 (Section 36(2)) (Health Research) Regulations 2018.

Inclusion Criteria:

• Birth at less than 32 weeks gestational age and admitted to the NICU
• Free from antenatal suspicion of major congenital abnormalities
• Ability of the participant (in the investigator's opinion) to comprehend the full nature and purpose of the study including possible risks and side effects.
• Consent to participate in the study and willing to comply with the protocol and study restrictions.

Exclusion Criteria:

• No informed consent
• Major congenital anomalies
• Gastrointestinal anomalies
• Suspected or confirmed inborn errors of metabolism

Infants admitted to the neonatal intensive care unit at Cork University Maternity Hospital, Ireland .