Computational protocol: Rapid Microbiome Changes in Freshly Deposited Cow Feces under Field Conditions

Similar protocols

Protocol publication

[…] Reads with a quality score less than 15 were trimmed using fastq_quality_trimmer from the FASTX toolkit version (). Forward and reverse barcode combinations were used to identify each sample. Paired-end reads were joined together using fastq-join from ea-utils.1.1.2-537 (), and unjoined reads were filtered out by an in house script. SeqFilters.jar was used to extract sequences between V4 region primers (), and chimeric reads were filtered out using UCHIME with the “Gold” database (). The final high quality reads were analyzed using QIIME version 1.8.0 under default parameters (). The operational taxonomic unit (OTU) picking was performed using UCLUST at a 3% threshold () and the OTUs with fewer than 10 reads were removed to avoid PCR sequencing errors. OTU taxonomy was assigned by RDP 2.2 at a confidence level cutoff of 80% (). The relative abundance (RA), as well as temperature and moisture reported here, is the averaged RA of two cowpats. U1, U2 and S1, S2 are the abbreviations for unshaded (U) and shaded (S) samples from Farm 1 and 2, respectively. Non-metric multidimensional Scaling (NMDS) based on the Bray–Curtis dissimilarity matrix was performed using the R-VEGAN package in R to visualize the community structure. Both replicate samples were plotted on NMDS to visualize their structural differences. Once the iterative NMDS solution is found, the R-VEGAN package uses a random permutation method (“envfit”) for determining whether factors and covariates are significantly influencing the NMDS scores. For categorical factors (unshaded and shaded fecal samples), “envfit” uses the difference between the average NMDS scores (centroids) at each level of the factor. Since significant differences were found among different factors (see “Results” section), envfit was also used to determine if there is a significant difference between two individual centroids. For a covariate (moisture), the test compares the length of the fitted vector on the NMDS axes to a distribution of fitted vector lengths calculated after randomly permuting the dataset. Ten-thousand permutations were run to construct the distributions of fitted vector lengths and factor-level differences. We defined a p-value (p) less than 0.05 as a significant difference for factors and covariates.Alpha diversity analysis was calculated via a rarefaction curve from QIIME’s alpha diversity pipeline (). Duplicate samples were combined and normalized to 10.5K reads, which was equivalent to the sample with the least reads, for estimating species richness (Chao1) and diversity (Shannon index). Evenness was calculated by dividing the Shannon index by the natural log of Chao1.In addition to community structure analysis, Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) was also used to infer the functional content, based on 16S OTU BIOM generated from QIIME (). The RA of PICRUSt inferred function is also reported here.To determine significant differences in RAs, simple t-tests were performed using R; a p-value (p) of ≤0.05 indicates a significant difference. […]

Pipeline specifications

Software tools ea-utils, UCHIME, QIIME, UCLUST, vegan, PICRUSt
Applications Phylogenetics, 16S rRNA-seq analysis
Organisms Bos taurus