3.2. Expression Comparisons of Fatty Acid (FA), Triacyl Glyceride (TAG) Biosynthesis, Citric Acid Cycle (TCA) and Glycolysis Genes between Custom Oil Palm Mesocarp, Arabidopsis and Rice Microarrays at 16 Week After Anthesis (WAA)
In the custom oil palm mesocarp microarray, probes designed to detect FA and TAG biosynthesis genes during mesocarp development were included. To compare the detection efficiencies of FA and TAG biosynthesis genes on the different microarray platforms, we compared the expression levels through signals produced from microarray hybridization in mesocarp tissue at 16 WAA, at which point mesocarp oil accumulation in this tissue is entering the exponential phase [39]. In the Arabidopsis and rice microarrays, the expression signals produced at 16 WAA were similar across the selected FA and TAG genes (Figure 2A), however up to 3-fold variation was observed in the oil palm microarray hybridization signals. The statistical test (Mann-Whitney test) showed that the oil palm custom array produced significantly higher (p-value < 0.01) probe signal intensity in all the FA and TAG probes when compared to signals produced using Arabidopsis microarray. Comparing the rice microarray and oil palm custom array, we also found a similar situation to the Arabidopsis-oil palm array comparison. The Mann-Whitney test also showed significant differences overall in the probe signal levels of transcript that code for FA and TAG genes using the oil palm microarray as compared to the rice microarray platforms, at p-value < 0.05. This was with the exception of the probe that coded for Carboxyltransferase α-subunit of heteromeric acetyl-CoA carboxylase (ACC CT-α), where no significant differences were seen between rice and oil palm arrays. (p-value, 0.66).
Figure 2  Expression signal comparison of selected fatty acid (FA) and triacylglycerols (TAGs). (A) tricarboxylic acid cycle (TCA) and glycolysis (B) genes at 16 Weeks After Anthesis (WAA) using different microarray platforms, Rice, Arabidopsis and Oil Palm. Signals were compared using a Mann-Whitney test at p-value < 0.05. FAD 2, Oleate desaturase; PK, Pyruvate kinase; LAC, Long-chain acyl-CoA synthetase; FAT A, Acyl-ACP thioesterase A; KAS I, Ketoacyl-ACP synthase I; KAS III, Ketoacyl-ACP synthase III; KAS II, Ketoacyl-ACP synthase II; SAD, Stearoyl-ACP desaturase; FAT B, Acyl-ACP thioesterase B; CPT, Diacylglycerol cholinephosphotransferase; DGAT 2, Acyl-CoA: diacylglycerol acyltransferase 2; LPCAT, 1-acyl glycerol-3-phosphocholine acyltransferase; KAR, Ketoacyl-ACP reductase; ACC CT-α, Carboxyltransferase α-subunit of acetyl-CoA carboxylase; HAD, hydroxyacyl-ACP dehydrase; WRI1, EAR, Enoyl-ACP reductase; PDAT, Phospholipid:diacylglycerol acyltransferase; GPAT, glycerol-3-phosphate acyltransferase; MAT, Malonyl-CoA:ACP malonyltransferase; LPAAT, Lyso PA acyltransferase; DGAT 1, Acyl-CoA:diacylglycerol acyltransferase 1; PDH (DHLAT), Dihydrolipoamide acetyltransferase; SCS, Succinyl coenzyme A synthetase; PGK, Phosphoglycerate kinase; IDH, Isocitrate dehydrogenase; PFK-1, Phosphofructokinase 1; CS, Citrate synthase; PGI, phosphoglucose isomerase; ALDOA, Fructose-bisphosphate aldolase; TPI, Triosephosphate isomerase; ENO 1, enolase 1; ME, Malic Enzyme; ACLY, ATP Citrate Lyase; GAPDH, Glyceraldehyde-3-phosphate dehydrogenase; MDH, Malate dehydrogenase; FH, Fumarase; HK, Hexokinase.    Similar results were also observed when comparing the glycolysis and TCA pathway genes (Figure 2B). Most of the genes exhibit higher expression signals in the oil palm microarray dataset and show significant differences in their expression (p-value < 0.01) when compared to the Arabidopsis array, with the exception of the transcript that coded for fumarase (p-value, 0.1689). In the Arabidopsis microarray, only the transcript coding for enolase 1 had significantly higher expression signal (p-value, 0.0004) compared to oil palm microarray. Comparisons between rice and oil palm arrays yielded similar results, with most of the genes showing significantly higher expression signal using the oil palm array, with the exception of the transcripts for hexokinase, fumarase, glyceraldehydes-3-phosphate dehydrogenase (GAPDH), and ATP citrate lyase (p-value > 0.05).
Probe signal comparisons identified several genes with expression signals with up to 3-fold difference between Arabidopsis and the oil palm microarray, similarly for the rice microarray. The exceptions were aconitase and succinyl-CoA synthetase alpha subunit; these two transcriptsexhibited much higher signal in the rice microarray compared to the Arabidopsis microarray. For genes in FA, TAG, TCA and glycolysis pathways (Figure 2), the 16 biological samples used in the microarray showed consistent differences between replicates. This argues for a lack of dynamic detection range in the Arabidopsis and rice microarrays, compared to the oil palm microarray even for genes in conserved metabolic pathways—A reflection of the lack of homology between the Arabidopsis and rice array probes used here compared with oil palm FA, TAG, glycolysis and TCA genes.