Whatman biometra thermocycler

To design our Custom PCR array we took all miRNAs included in the Qiagen Human miRNome^™ miscript v. 16.0 PCR array and aligned them (using BLAST, [32 (link)]) to cow miRNA sequences listed in miRBase 19 (accessed 20/02/2013, [33 (link)]) in a Linux environment. A total of 308 miRNAs conserved in cow (i.e., with ≤ 2 nucleotide mismatches between human and cow sequences) were included in the 384-well Custom miScript miRNA PCR array (384-well, Qiagen).
Three sample pools from each of Day 0, Day 8 and Day 16 (3–4 samples / pool) were analysed. cDNA (10 μL) was synthesised from 2 μL of RNA sample using miScript II RT kit (Qiagen) in a Whatman-Biometra Thermocycler (Biometra, USA). The arrays were setup according to the manufacturer’s instructions and were analysed on the LightCycler 480 System (Roche, Switzerland). Data analysis was performed using Microsoft Excel (Microsoft Corporation, USA) and R programming using R language 3.02 and RStudio 0.98 [30 , 31 ]. Raw Cq data were initially filtered to remove wells with non-specific amplification as identified by melting-curve analysis. Cq values were normalised using the global mean expression, which was calculated from miRNAs which were detected in all of the sample pools. The statistical analysis of the transformed normalised data was performed as described for the sequencing data above. The PCR array dataset is provided in S2 File.

We used commercial 384-well Custom PCR arrays (Qiagen) which covered a total of 377 unique human miRNAs, 308 of which were conserved in cow (2 nucleotide mismatch allowed). Three pools of 3 to 4 samples each, from each of Day 0, Day 8 and Day 16 for non-pregnant animals, and from Day 24 pregnant animals (12 pools in total) were analysed. cDNA (10 μL) was synthesised from 2 μL RNA sample using miScript II RT kit (Qiagen) in a Whatman-Biometra Thermocycler (Biometra, USA). The arrays were setup according to the manufacturer’s instructions and were run on the LightCycler 480 System (Roche, Switzerland). Data analysis was performed using Microsoft Excel and R programming. Raw Cq data were initially filtered to remove wells with non-specific amplification as identified by melting-curve analysis. Next, miRNAs which had Cq > 35 in more than 75 % of samples per experimental group were removed from the dataset. Cq-values were normalised using the global mean expression, which was calculated from the miRNAs which were detected in all of the sample pools. The mean expression across non-pregnant time-points (Days 0, 8 and 16) was used for analyses, as for the sequencing data, resulting in 3 data-points for analyses from each of pregnant and non-pregnant groups (NP vs P24). The statistical analysis of the transformed normalised data was performed as described for the sequencing data above.