Infants, aged 8–14 months and starting daycare within 12 weeks after intervention start, were included in the ProbiComp study , a double blinded randomized placebo-controlled intervention. BB-12® and LGG® are registered trademarks of Chr. Hansen A/S who kindly provided the study products. During the 6 months intervention period participants were randomized to receive either a combination of the two probiotics BB-12® and LGG® at a dose of 109 colony-forming units (CFU) each, or a placebo (maltodextrin), during the winter season. Participants were instructed not to consume food products and supplements containing probiotics 2 weeks prior to and during the intervention period. The study spanned two consecutive winter seasons and included a total of 290 infants (145 for each season). Study population characteristics (Table 1) were obtained through parental interviews.
Fecal samples, DNA extraction and PCR amplification of theV3 region of the 16S rRNA gene
Fecal samples obtained before (N = 265, age 8–13 months) and after (N = 210, age 14–19 months) the intervention were freshly delivered on the morning of the visit or had been stored in the participant’s home, either in the freezer (−18 °C) or in the fridge (4 °C) for maximally 24 h before delivery to the University of Copenhagen, Department of Nutrition, Exercise and Sports, where they were stored at −80 °C until DNA extraction. Extraction was done in random order, but the two samples obtained from the same child were always processed together (N = 16–20 samples/extraction round). 250 mg feces was applied and treated according to the protocol provided by the manufacturer (PowerLyzer® PowerSoil® DNA isolation kit, MoBio 12,855–100) with minor modifications: Bead beating was performed at 30 cycles/s for 10 min (Retsch MM 300 mixer mill) and the initial centrifugation steps were performed at 10,000 x g for 3 min, as recommended for clay matter. DNA quantity and quality was measured by NanoDrop® 1000 (ThermoFisher Scientific), yielding on average 85.0 ± 47.1 ng/μl DNA with A260/A280 = 1.89 ± 0.10 and A260/A230 = 1.71 ± 0.39. The PCR amplification of the V3-region of the 16S rRNA gene was performed with 5 ng community DNA as template, using 0.2 μl Phusion High-Fidelity DNA polymerase (Fisher Scientific, F-553 L), 4 μl HF-buffer, 0.4 μl dNTP (10 mM of each base), 1 μM forward primer (PBU 5′-A-adapter-TCAG-barcode-CCTACGGGAGGCAGCAG-3′) and 1 μM reverse primer (PBR 5′-trP1-adapter-ATTACCGCGGCTGCTGG-3′) in 20 μl total reaction volume. Both primers included sequencing adaptors and the forward primer additionally a unique 10–12 bp barcode (Ion Xpress™ Barcode Adapters). PCR amplification included 30s at 98 °C, 24 cycles of 15 s at 98 °C and 30s at 72 °C, followed by 5 min at 72 °C. PCR products were purified by use of HighPrep™ PCR Magnetic Beads (MAGBIO®, AC-60005) with the 96-well magnet stand (MAGBIO®, MyMag 96), according to the manufacturers recommendations. DNA quantity was measured using Qubit® dsDNA HS assay (Invitrogen™, Q32851) and a total of 456 samples contained sufficient DNA for sequencing. Samples were pooled to obtain equimolar libraries containing up to 86 samples (randomized across treatment groups and age) in each library.
DNA sequencing and data handling
Sequencing of the 16S rRNA amplicon libraries was performed using the Ion OneTouch™ and Ion PGM platform with a 318-Chip v2, generating 4–5 million reads per chip with a median length of 180 bp. Sequencing data were imported into CLC Genomic Workbench (version 8.5. CLC bio, Qiagen, Aarhus, DK) and using the NGS core tools, reads were assessed for quality (QC report), de-multiplexed according to barcode (Demultiplexing) and trimmed (Trim sequences) to remove barcodes and PCR primers (minimum alignment score 17/17, discard when both primers were not found) and to discard reads below 125 bp and above 180 bp. Quality filtering (−fastq_filter, maxee 1.0), dereplication (−derep_fulllength), OTU clustering (−cluster_otus, minsize 4), chimera filtering (−uchime_ref, RDP_gold database), mapping of reads to OTUs (−usearch_global, id 97%) and generation of OTU table (python, uc2otutab.py) was done according to the UPARSE pipeline , generating a total of 1096 non-chimeric OTUs (Operational Taxonomic Units). In QIIME , the OTU table was filtered (filter_otus_from_otu_table.py/filter_fasta.py) to exclude OTUs with average relative abundance below 0.005% of the total community, resulting in 330 OTUs. Taxonomy was assigned (assign_taxonomy.py), using the rdp classifier with confidence threshold 0.5 (recommended for sequences shorter than 250 bp ) and the GreenGenes database v. 13.8 . In addition, the taxonomy of selected OTUs was confirmed/disconfirmed by BLAST  search against the 16S rRNA sequence database. Based on PyNAST alignment of representative OTU sequences (align_seqs.py, filter_alignment.py, default settings), including an archaea (Methanosarcina) full length 16S rRNA gene sequence as outgroup for rooting, a phylogenetic tree was created with FastTree (make_phylogeny.py, default settings) and re-rooted to the outgroup, which was subsequently pruned, using Dendroscope v3.5.7 . Taxon abundances (average classification: Phylum: 100%, Family: 99.9%, Genus: 91.8%) and alpha diversity (Shannon index, Observed OTUs, Pielou’s evenness index) and beta diversity (UniFrac distances) were calculated in QIIME (summarize_taxa.py, alpha_rarefaction.py, beta_diversity_through_plots.py), with the sequencing depth rarefied to 10,000 sequences per sample for diversity analysis. OTU abundances were estimated by total sum scaling.
Detection of the probiotic strains BB-12® and LGG®
Primer design, PCR and gel electrophoresis
Primers specific for the probiotic strains were designed using the primer3 online software tool (http://bioinfo.ut.ee/primer3-0.4.0/). Selection of primers for the LGG® strain were based on a phage-related gene target sequence  previously reported to be strain specific , whereas primers for the BB-12® strain were selected within the CRISPR-Cas system , specifically the Type I-U CRISPR-Cas7 gene in BB-12® (for primer sequences see Additional file 1: Table S1). Each PCR reaction contained 14.92 μl DNAse-free water, 2 μl 10X Accuprime PCR buffer II, 1 μl forward primer (0.5 μM final conc.), 1 μl reverse primer (final conc. 0.5 μM), 0.08 μl AccuPrime Taq Polymerase (Invitrogen, 12,346–086) and 1 μl template DNA (1 ng/ul) in a final volume of 20 ul. The PCR program (Veriti Thermal Cycler, Applied Biosystems™, 4,452,300) included 2 min at 94 °C, 35 cycles of 20 s at 94 °C, 20 s at 65 °C and 30 s at 68 °C, followed by a final extension for 5 min at 68 °C. The PCR products were separated on a 1.5% agarose gel with 1:10,000 volume SYBR safe (Bio-Rad) for 1 h at 100 V before imaging (Molecular Imager® GelDoc™ XR System).
Primer validation and probiotic quantification
To validate primer specificity, a selection of Lactobacillus and Bifidobacterium strains were cultivated on MRS (de Man, Rogosa and Sharpe, ) or BSM (Sigma, 88,517) agar at 37 °C under anaerobic conditions for 72 h (Additional file 1: Table S2). Genomic DNA was extracted (PowerLyzer® PowerSoil® DNA isolation kit, MoBio 12,855–100) from colony material according to the protocol provided by the manufacturer and diluted to 1 ng/μl and used for input in the PCR. For spiking fecal samples with known quantities of the strains LGG® and BB-12®, single colonies of these were propagated in 10 ml MRS medium under anaerobic conditions for 48 h at 37 °C, centrifuged at 10,000×g for 5 min at 4 °C, re-suspended in 1 ml maximum recovery diluent (MRD, Oxoid CM0733), and 10-fold serially diluted in MRD. One hundred μl of each dilution (100–10−7) was spiked into 100 mg LGG®/BB-12® negative baseline fecal samples (based on specific qPCR of the probiotic strains) and also plated in duplicates (dilutions 10−4–10−9) on MRS plates and incubated anaerobically at 37 °C for 24 h before CFU counting. Community DNA was extracted (PowerLyzer® PowerSoil® DNA isolation kit, MoBio 12,855–100) from each spiked samples plus none-spiked controls.
Each quantitative PCR (qPCR) reaction contained 5 μl PCR-grade water, 1 μl forward primer (final conc. 0.5 μM), 1 μl reverse primer (final conc. 0.5 μM), 10 μl SYBR Green I Master 2X (LightCycler® 480 SYBR Green I Master, Roche, 04887352001) and 2 μl template DNA, in a 20 μl total volume. Template DNA was either 10-fold serial dilutions of 1 ng/μl DNA extracted from pure cultures of LGG® or BB-12®, community DNA isolated from fecal samples (LGG®/BB-12® negative) spiked with 10-fold serial dilutions of LGG® or BB-12® cultures, or community DNA (5 ng/μl) extracted from fecal samples obtained during the intervention. Triplicate samples were run on the qPCR instrument (LightCycler® 480 Instrument II, Roche, 05015243001) and the program included 5 min pre-incubation at 95 °C, followed by 45 cycles with 10 s at 95 °C, 15 s at 65 °C and 15 s at 72 °C. A melting curve analysis was subsequently performed with 5 min at 95 °C, 1 min at 65 °C and continuous temperature increase (ramp rate 0.11 °C/s) until 98 °C. Data were analyzed with the LightCycler® 480 Software (v 1.5) and levels of the strains LGG® and BB-12® in samples were quantified based on standard curves generated from extracted community DNA of fecal samples spiked with 10-fold serial diluted DNA extracted from pure cultures of LGG® or BB-12®. Samples were regarded LGG® and/or BB-12® positive when at least 2 of the 3 replicates were positive, the melting curves were consistent with that of pure cultures of the respective strains, and Ct values were within the detection range achieved for the 10-fold dilution series standard-curves.
Statistics were performed with the GraphPad Prism software (v. 7.0, GraphPad Software Inc., La Jolla, CA), using the Spearman’s rank correlation, Mann-Whitney test and Fisher’s exact test or by use of specific scripts for analyzing sequencing data as implemented in QIIME v. 1.9 . For comparison of beta diversity the compare_categories.py (adonis, permutations = 999) script was used and for comparison of microbial composition at different taxonomical levels the group_significance.py (mann_whitney_u, Benjamini-Hochberg False Discovery Rate (FDR) correction , threshold 0.05) script was applied. Alpha diversity measures were compared between groups using the Mann-Whitney test of the average of 10 iterated values obtained following rarefying to 10,000 reads per sample.