In order to identify genes that are differentially expressed in response to factors such as long-term cold storage (C), genes that are differentially expressed during ripening (R), as well as the combined effects of these two factors (long-term cold storage and ripening), a 2 × 2 factorial design was used to generate four cDNA libraries that can be used for statistical comparative analyses [
16].
These libraries represent the mesocarp tissue of O'Henry peaches (Prunus persica) under the following four distinct post-harvest conditions:
E1 = non-ripe; no long-term cold storage = R-, C-
E2 = ripe; no long-term cold storage = R+, C-
E3 = non-ripe; long-term cold storage = R-, C+
E4 = ripe: long-term cold storage = R+, C+
The maturity and physiological parameters of the fruits used to create these libraries are reported in Campos-Vargas et al [
3]. These maturity and physiological parameters include firmness, percentage of total soluble solids, respiration rate and ethylene production [
3].
5' end sequencing of clones from these libraries resulted in a total of 50,625 EST sequences. Assessment of these EST sequences revealed that a total of 41,519 ESTs are "good quality" sequences (average Phred Q>20 between bases 100 and 300 for each EST). Relatively equal numbers of "good quality" ESTs from each of the four post-harvest conditions were used for further analyses (Table , Figure , Additional File
1, Table S1).
| Table 1Prunus persica Unigene Set Statistics |
As of May 2009, there were 79,023
Prunus persica EST sequences in NCBI, of which 32,497 ESTs correspond to sequences reported in this manuscript, but were publicly release to NCBI in 2006 (accession numbers: DY633390-DY654328, DW347789-DW359346). These 32,497 pre-release EST sequences also make up almost half of the sequences available in the ESTree database (75,404 ESTs) [
17,
18]. In addition to the sequences we have release previously, we are releasing an additional 9,022 EST sequences with the publication of this manuscript. This is a total of 41,519 ESTs which are being reported for the first time in this manuscript.
Assembling these 41,519 ESTs using CAP3 95/60 resulted in the identification of 10,830 Unigenes (4,169 contigs and 6,661 singletons). EST distribution in the contigs is presented in Additional File
1, Table S2. BLASTn and BLASTx analyses revealed that 98.9% of the contigs shared homology to sequences in public databases, whereas the remaining 47 contigs and 656 singletons showed no significant homology to the sequences in the public databases. Interestingly, although there is a high level of homology between the consensus sequences of our Unigenes and the sequences in public databases, 3,786 of our Unigenes contain novel sequence information, not present in public databases such as NCBI, ESTree, PlantTA, GDR and ChillPeach [
13,
17-
21].
BLAST, Interproscan and BLAST2GO analyses were used to annotate the Unigenes and ESTs [
22-
26]. Gene Ontology (GO) annotations were assigned to the ESTs and contigs. Detailed information associated with the ESTs, contig consensus sequences and their associated annotations are available to browse or download at
http://www.genomavegetal.cl/juice/EST[
27].
The length of the contig consensus sequences ranged from 0.1 to 3.0 kb, with an average length of 0.9 kb (Table ). This large size detected in the consensus sequences suggests that some of these cDNA clones may be full-length. Analyses of the contig consensus sequences with EuGène'Hom [
28] enabled us to identify the components of a full-length cDNA (5' UTR, ORF and 3' UTR). These analyses identified 614 full-length cDNAs, ranging in size from 0.4 to 2.7 kb (Table ). Additionally, 1,109 contigs represent putative full-length cDNAs, in which the 5' UTR and ORF regions were detected. Since we performed 5'-end sequencing on these putative full-length cDNAs clones, they should contain the 3' UTR. Further 3' sequencing of these clones will confirm the full-length clones.
| Table 2Full-length cDNA sequences identified with EuGène'Hom. |
A summary of the results from this EuGène'Hom analysis is presented in Table . The clones of these full-length cDNAs have been identified by correlating the ESTs that are at the 5' end of the consensus sequences with the clones that these ESTs represent. Correlations between the insert size of many of these clones and the size of the contig consensus sequence suggests that the EuGène'Hom predictions are accurate (data not shown). Since our cDNA libraries were made in Clontech's pDNR-1r vector, these full-length and putative full-length cDNAs have flanking loxP recombination sites on both sites of the cDNAs [
29,
30].
Differential expression of genes in each post-harvest stage
To estimate the distribution of ESTs from individual contigs among four cDNA libraries, a two-dimensional hierarchical clustering using pair-wise average-linkage cluster analysis [
31] was applied. To ensure that contigs made up of a low number of ESTs are not unrepresented in the cluster analyses, the number of ESTs in each library was normalized proportionally to the total number of ESTs in their corresponding contig. The hierarchical clustering of the normalized contigs (Additional File
1, Figure S1), revealed that the distribution of ESTs in E1 (R
-, C
-) and E3 (R
-, C
+) are more similar to each other when compared to E2 (R
+, C
-) or E4 (R
+, C
+). Similarly, E2 (R
+, C
-) and E4 (R
+, C
+) are more similar to each other when compared to E1 (R
-, C
-) or E3 (R
-, C
+).
Relative gene expression can be determined by statistically comparing the number of ESTs from a gene between different libraries or different genes in the same library [
32]. Pair-wise comparisons of the number of ESTs from each post-harvest condition, revealed differentially expressed genes (p < 0.01). Based on this statistical analysis, 197 contigs with at least one significant difference in gene expression were detected between the different post-harvest conditions. A total of 30 genes increase expression in ripe, juicy fruits (E2: R
+, C
-) when compared to non-ripe fruits (E1: R
-, C
-), whereas 39 genes are reduced (Additional File
2, Table S4). Long-term cold stored fruits (E3: R
-, C
+) increases the expression of 56 genes and decreases the expression of 45 genes when compared to fruits that have not undergone this long-term storage (E1: R
-, C
- ; Additional File
2, Table S5). Fruits that are ripened following long-term cold storage and are woolly (E4: R
+, C
+) have an increased expression of 36 genes and a decreased expression of 68 genes when compared to unripe fruits that have undergone long-term storage (E3: R
-, C
+; Additional File
2, Table S6). A total of 9 genes have a statistically significant increase in expression in woolly ripened long-term cold stored fruits (E4: R
+, C
+) when compared to juicy ripened fruits that have not been stored in the cold (E2: R
+, C
-), whereas 8 genes are reduced (Additional File
2, Table S7).
This
in silico analysis was validated by performing qPCR analyses on several of the contigs that showed differential gene expression levels based upon the pair-wise
in silico analyses (Figure , Additional File
2 Tables S4-S7). Figure reveals differential gene expression confirmed by performing qPCR analyses of the following contigs under the four post-harvest conditions: Wcor (contig30), dormancy-associated protein (DRM1; contig438), polygalacturonase (contig 1123), luminal binding protein 1 (BIP-1; contig2715), temperature-induced lipocalin (TIL) (contig 2980), and lipoxygenase (contig 3870) (Figure ). This comparative analysis revealed that the
in silico and qPCR assays yield similar results for these genes, under the conditions analyzed.
Ripening related genes
The pair-wise comparison of EST abundance of our contigs between two stages, revealed a significant increase in the transcript levels of 30 genes in these ripe fruits (E2: R
+, C
-) when compared to the unripe fruits (E1: R
-, C
-), and a reduction in the transcript levels of 39 genes (Additional File
2, Table S4). It is important to note that the fruits reported by Campos-Vargas et al [
3] are the same fruits that were used for constructing our libraries. As reported by Campos-Vargas et al [
3], ripe fruits were less firm than unripe fruits (7.9 N versus 72.5 N, respectively), but had a higher respiration rate (110.8 versus 28.8 mL CO
2 kg
1 h
1, respectively) and ethylene production (4.7 versus 1.6 μL C
2H
4kg
1h
1) [
3,
33-
38].
The annotations of these differentially expressed transcripts reveal genes that participate in the physiological changes detected between ripe and unripe fruits. For example, 5 polygalacturonase transcripts (contigs1123, 1507, 1745, 3321, and 3904) are accumulated higher in ripe fruits than unripe fruits (Additional File
2, Table S4), which may lead to the changes in fruit firmness. These contigs share high identity (97-99%) with the polygalacturonase that was identified in microarray analyses of ripening fruits [
11,
39]. These results are consistent with the increase of endopolygalacturonase enzymatic activity in ripened peach fruits which has been reported previously [
40,
41].
In addition to the increased transcript accumulation of polygalacturonases, we detected a decreased expression of pectate lyase (contig 3683) and pectin methylesterase inhibitor genes (contigs 59 and 3598) (Additional File
2, Table S4). Trainotti et al [
11] have detected increases in pectate lyase transcripts during the S4I phase of peach fruit ripening (climateric fruits that have not undergone softening, 115 days post-fertilization) and then decreases in the S4II phase (climateric fruits that have softened, 120-125 days post-fertilization). Our pair-wise comparisons of ESTs between ripe and unripe fruits detect a decrease in pectate lyase in ripe fruits when compared to unripe fruits (Additional File
2, Table S4). This decrease in pectate lyase gene expression as well as the reduction in firmness of the ripened fruits [
3] suggests that our E2 stage is similar to the S4II stage described by Trainotti et al [
11].
In addition to cell wall modifying enzymes, we have detected increased expression of genes associated with ethylene biosynthesis such as S-Adenosyl Methionine Synthetase (SAM Synthetase, contig 3949) and 1-Aminocyclopropane-1-Carboxylate Oxidase (ACC oxidase, contig 1901) (Additional File
2, Table S4). The increase in expression of this ACC oxidase (contig 1901) in ripe fruits has also been detected by Trainotti et al [
39], whereas the SAM synthetase was not. The differential expression of these genes may also explain the increase in ethylene production detected in the ripe (E2: R
+, C
-) fruits.
As mentioned earlier, we have also detected an increased respiration in ripe fruits when compared to unripe fruits [
3]. The pair-wise comparison reveals increased expression in several genes associated with glycolysis which may lead to this increase in respiration (Additional File
2, Table S4). These genes include Fructose-bisphosphate aldolase (contig 1358), glyceraldehyde 2-phosphate dehydrogenase (contig 1653) and pyruvate decarboxylase (contigs 1054 and 2749).
Cold-storage related genes
Changes in firmness and respiration rate were also seen when unripe fruits stored at 4°C (E3: R
-, C
+) were compared to unripe fruits prior to storage (E1: R
-, C
-; firmness: 43.2 versus 72.5 N; Respiration rate: 94.3 versus 28.8 CO
2 kg
1 h
1, respectively) [
3]. The pair-wise analyses reveal 56 genes that increase expression and 45 genes that decrease expression as a result of this long-term cold storage (Additional File
2, Table S5). Several of these genes are associated with cell wall organization and, therefore, may be associated with the changes detected in firmness. Genes that increase include: polygalacturonase inhibiting protein 1 (contig 2988), pectin methylesterase (contig 2877), basic endochitinase (contig 2131), and endopolygalacturonases (contigs 1507 and 1745). Genes that decrease include: expansin (contig 1200), invertases/pectin methylesterase inhibitors (contigs 59, 1945 and 3598), and pectate lyase (contig 3683). Among these cell wall associated contigs, only the basic endochitinase (contig 2131) has been reported previously to have altered expression in long-term cold stored fruits [
13].
In addition to the altered expression of cell wall modifying proteins, we have also detected increased expression of glutathione peroxidase (contig 2514; Additional File
2, Table S5). Glutathione peroxidase is an oxidative stress responsive gene, which suggests that the fruits that have been stored for prolonged periods of time have an increase in oxidative stress, possibly associated with the increased respiration rates. Oxidative stress has been demonstrated in many species to be associated with chilling injury [
42-
45].
Among the genes differentially expressed between fruits that have undergone prolonged cold storage with those that have not, seven of these correspond to the genes that Ogundiwin et al [
13] have also identified as having altered expression following cold storage (contigs 1546, 1693, 2131, 2311, 2495, 2957 and 2988). The remaining 94 contigs that are differentially expressed between unripe fruits stored at 4°C (E3: R
-, C
+) and unripe fruits prior to storage (E1: R
-, C
-) are novel and may participate in the adaptation and/or stress responses associated with long-term cold storage of these fruits.
Wooliness response genes
When fruits that have been stored for long periods of time in cold storage (E3: R
-, C
+) are removed from this storage and ripened (E4: R
+, C
+), there is a significant decrease in firmness (43.2 versus 6.8 N, respectively), a significant increase in ethylene production (1.0 versus 18.1 μL C
2H
4kg
1h
1, respectively) and the fruits have a woolly phenotype [
3]. The pair-wise analyses reveal an increased expression of 36 genes and a decreased expression of 68 genes in fruits that are ripened following long-term cold storage and are woolly (E4: R
+, C
+) in comparison to unripe fruits that have undergone long-term storage but are unripe (E3: R
-, C
+; Additional File
2, Table S6). Additionally, these analyses reveal an increased expression of 9 genes and decreased expression of 8 genes when woolly fruits are compared to juicy fruits, (E4: R
+, C
+ and E2: R
+, C
-, respectively; Additional File
2, Table S7).
The changes in firmness and the wooliness phenotype have been reported previously to be correlated with modifications in the levels of cell wall modifying enzymes [
1,
9,
40,
41,
46-
49]. As mentioned earlier, our pair-wise analyses have identified five polygalacturonase transcripts with increased accumulation in ripe fruits when compared to unripe fruits (contigs 1123, 1507, 1745, 3321 and 3904; Additional File
2, Table S4-7). Of these five transcripts, two (contigs 1507 and 1745) have an increased accumulation following long-term cold storage. Ripening of the cold stored fruits produces an increased accumulation of four of these polygalacturonase transcripts (contigs 1507, 1745, 3321 and 3904). However, when comparing juicy fruits with woolly fruits, there is less accumulation of three of these transcripts (contigs 1123, 1507 and 1745) in woolly fruits. These results suggest that in woolly peaches there are diverse polygalacturonase transcripts which are unable to accumulate to the levels normally present in ripe juicy fruits. These results are complementary to several studies that have linked polygalacturonase activity with wooliness [
1,
9-
12,
39-
41,
46-
50]. Interestingly, Peace et al [
51] have also identified an endopolygalacturonase that co-localizes with a major QTL affecting both mealiness and bleeding. The large number of different polygalacturonase transcripts may be a result of alternative splicing or a recent gene duplication resulting in a highly conserved polygalacturonase gene family. The completion of the peach genome sequence will help to clarify whether these transcripts are a result of multiple genes or alternative splicing.
In addition to the changes in the accumulation of polygalacturonase transcripts, we have also identified a polygalacturonase inhibiting protein that has an inverse expression when compared to polygalacturonase in fruits that are ripened following a long-term cold storage and have a woolly phenotype. That is, there is a reduced expression of the polygalacturonase inhibiting protein in woolly fruits (E4: R+, C+) when compared to unripe fruits that have undergone long-term cold storage (E3: R-, C+). This reduction in the expression of the polygalacturonase inhibiting protein may lead to an increase in polygalacturonase activity and subsequent reduction in the fruit firmness. A similar pattern of reduced expression in fruits that are ripened following a long-term cold storage is also seen for another cell wall modifying enzyme, endochitinase (contig 2131).
A decrease in pectate lyase expression (contig 3683) was detected when comparing ripen fruits E2 (R
+, C
-) with unripe fruits E1 (R
-, C
-). However, a decrease in expression of this contig was also detected in fruits stored for prolonged periods of time in the cold (E3: R
-, C
+ versus E1: R
-, C
-), suggesting that this decrease may be similar to the decrease detected by Trainotti et al [
11] in the S4II stage (climateric fruits that have softened, 120-125 days post-fertilization). Interestingly, the expression levels of this pectate lyase increased in woolly ripened long-term cold stored fruits (E4: R
+, C
+) when compared to long-term cold stored fruits (E3: R
-, C
+). We have previously reported a decreased transcript level of pectate lyase when woolly fruits were compared to juicy fruits by macroarray analyses [
12]. One explanation for this is that the cold-storage may have reduced the pectate lyase transcript to a point that despite the increased expression of pectate lyase in woolly ripened long-term cold stored fruits (E4: R
+, C
+) when compared to long-term cold stored fruits (E3: R
-, C
+), the transcript levels in woolly fruits (E4: R
+, C
+) are still lower than the levels detected in juicy fruits (E2: R
+, C
-). Alternatively, there may be a family of pectate lyase genes that are differentially expressed in the fruits, postharvest. It has been reported that in numerous species, pectate lyases are encoded by a large gene family [
52]. We have identified three putative pectate lyases among our Unigenes (contigs 328, 1989 and 3683), suggesting that in peaches there is a pectate lyase gene family.
Trainotti et al [
11] have demonstrated that three different expansin genes are expressed differentially during ripening. Two of these expansin genes are up-regulated and one is down regulated in the S3/S4 transition [
11]. We have identified seven contigs that are annotated as expansin or expansin family proteins (contigs 618, 1200, 1520, 1680, 2060, 2212 and 2420). The expansin genes analyzed by Trainotti et al [
11] correspond to contigs 3786 and 1951 in our unigene sets. Interestingly, we have identified a novel expansin (contig 2212) that has increased transcript levels in woolly fruits when compared to juicy fruits. Since there are higher levels of the transcript for this contig in ripe, juicy fruits (E2: R
+, C
-) in comparison to unripe fruits that have been stored for a long period of time (E3: R
-, C
+), the increase in expansin levels in woolly fruits (E4: R
+, C
+) is occurring during the ripening process by a factor that cold storage has affected, and is not due to an increase in expansin levels in cold-stored fruits.
Trainotti et al [
11] have also seen an increased expression of pectin methylesterase in S4I (climateric fruits that have not undergone softening, 115 days post-fertilization) and S4II ripened fruits (climateric fruits that have softened, 120-125 days post-fertilization). Our pair-wise comparison has detected an increased pectin methylesterase-like protein transcript (contig 2877) in woolly fruits when compared to juicy fruits. Increased transcripts that correspond to this contig were detected when comparing long-term cold storage fruits (E3: R
-, C
-) with fruits from packing (E1: R
-, C
-). This suggests that the increase in this pectin methylesterase gene was induced by cold storage and remained elevated following ripening and the development of the woolly phenotype.
In addition to the changes in the expression of pectin methylesterase, we have detected eight pectin methylesterase inhibitor proteins in our Unigenes (contigs 59, 422, 741, 1062, 1945, 2574, 3374 and 3598). Pair-wise analyses detected differential expression of three of these contigs (contigs 59, 1945 and 3598). Two of these contigs (contigs 59 and 3598) have similar patterns of differential expression under different post-harvest conditions. Both of these contigs have a reduced expression in ripe fruits with and without long-term cold storage (Additional File
2, Table S4-7). Contig 1945 also has a reduced expression in ripe juicy fruits. However, this contig has an increased expression in woolly ripened long-term cold stored fruits (E4: R
+, C
+) when compared to unripe fruits that have undergone long-term storage (E3: R
-, C
+; Additional File
2, Table S6). In the model plant Arabidopsis, there is a family of pectin methylesterase inhibitors and invertases [
53]. In vitro studies of pectin methylesterase inhibitors have been shown to have specific targets [
53]. Therefore, the differential expression between the different contigs that code for pectin methylesterase inhibitors may have different targets. However, it is important to mention that these different contigs may be the product of different transcripts from a single locus, rather than a gene family. Future mapping of these ESTs to the peach physical map will help to clarify this possibility [
54].
As mentioned earlier ethylene production has been shown to participate in peach fruit ripening [
10,
39]. We have detected an ACC oxidase gene (contig 1901) whose expression is increased in ripe fruits when compared to unripe fruits. However, there is a reduction in ACC oxidase (contig 1901) expression in woolly fruits (E4: R
+, C
+) when compared to juicy fruits (E2: R
+, C
-). This reduced expression of ACC oxidase may be associated with the wooliness phenotype. However, the ethylene production in the woolly fruits was higher than juicy fruits [
3], suggesting that there are other genes associated with ethylene production that may also be altered.
We have also identified a number of genes associated with membrane formation and stress related genes that are differentially expressed under the different postharvest conditions analyzed. A large number of the genes that we have identified as being differentially expressed in response to cold, ripening or a combination of these two stimuli, have not been reported previously. Therefore, the large number of genes identified as having altered expression in response to the individual or combined effects of ripening and long-term cold storage, demonstrates the complexity of the processes that are occurring in peach fruits post harvest and the differential expression of gene families. Further analyses of the expression of these genes and the factors that cause this differential expression may provide interesting clues about the mechanisms associated with wooliness.
Identification of genes that are co-expressed under different post-harvest conditions
To identify clusters of genes that are co-expressed under different post-harvest conditions, a
k-means analysis was performed to identify groups of genes with similar expression profiles.
K-means clustering was performed using 13 different clusters with a maximum of 50 iterations, each containing between 78 and 161 genes (Figure ). The optimal number of clusters was determined to be 13 by figure of merit analyses (Additional File
1, Figure S2).
Groups of clusters that correlate the gene expression profiles with the different post-harvest conditions are clearly visible when these groups are analyzed and compared (Figure , , and ). One group (Figure ) is represented by two clusters of genes (Clusters 4 and 12) whose expression profiles are inversely related in juicy, ripe fruits (E2: R
+, C
-), when compared to the other post harvest stages. Cluster 4 contains 107 genes with reduced expression exclusively in juicy, ripe fruits (E2: R
+, C
-), whereas cluster 12 contains 112 genes with increased expression exclusively in these same fruits. It is important to note this comparison is not a pair-wise comparison, rather a comparison between all four post-harvest conditions (E1: R
-, C
-; E2: R
+, C
-; E3: R
-, C
+; E4: R
+, C
+). Clusters 4 and 12 reveal a decrease and increase in gene expression in juicy fruits, respectively, which is not present in woolly fruits. Genes that are differentially expressed in juicy fruits and not woolly fruits are genes that may be participating in the woolly phenotype. Analyses of these sequences and their corresponding Gene Ontology annotations reveal that ripe peaches have an increase in metabolic processes, especially processes associated with the metabolism of carbohydrates, nucleic acids, amino acids, lipids and secondary metabolites (Figure ). In contrast, there is a decrease in translation, transport, and signal transduction processes. According to the GO annotations, it appears as though there is an increase in the expression of genes that are associated with mitochondria and plastids, as well as a decrease in those associated with ribosomes and the plasma membrane. Many of the metabolic process described earlier are carried out in the mitochondria and plastids, supporting this trend. The decrease in translation also correlates with the decrease in processes associated with the ribosome. These analyses suggest that woolly fruits lack the increased boost of metabolic processes necessary for ripening, such as the expression of cell wall modifying enzymes as well as the production of sugars and aromatic secondary metabolites. The changes in expression of cell wall modifying enzymes is consistent with what has been reported previously [
1,
9-
13,
39,
48,
51,
55,
56]. However, these cluster analyses reveal the organelles that may actively participate in these processes as well as a large number of genes and gene families that should be analyzed for their participation in the altered metabolic processes present in woolly fruits.
In addition to the absence of the modified gene expression patterns seen in clusters 4 and 12, there is another pair of clusters that represent the opposite tendency. Clusters 1 and 9 (Figure ) represent genes that change their expression exclusively in woolly fruits E4 (R
+, C
+). Cluster 1 contains 94 genes with increased expression exclusively in woolly fruits, whereas cluster 9 contains 95 genes with decreased expression exclusively in woolly fruits. In these clusters, there is a notable increase in the expression of genes associated with processes such as translation, transport, response to abiotic stress and cellular organization and biogenesis. Interestingly, the combined results of clusters 4 and 1 (Figures and ) demonstrate that woolly fruits have lost the reduced expression of genes associated with transcription and translation. Additionally, following abiotic stress conditions such as long-term cold storage, it is not surprising to see increased expression of abiotic stress related genes in woolly fruits. However, it should be noted that this class of abiotic stress related genes are distinct from those identified exclusively in the cold (Figure ). As was seen in clusters 4 and 12, many of the differentially expressed genes that are differentially expressed exclusively in woolly fruits appear to be associated with the mitochondria, plastids and ribosomes. The participation of the mitochondria and plastids also correlate well with the increase in respiration that was detected in these fruits [
3]. The plasma membrane is also participating, possibly associated with abiotic stress responses such as cold and/or desiccation stresses.
The identification of genes that are associated with mitochondrial and plastid processes, as well as the differential expression of genes such as glutathione peroxidase, suggests that oxidative stress and metabolic imbalances may play a major role in wooliness. Interestingly, recent publications have presented physiological analyses that support this hypothesis [
57-
59].
Yet another group represents genes that are differentially expressed exclusively in unripe fruits that have undergone long-term cold storage E3 (R
-, C
+; Figure ). This group is represented by clusters 2 and 6. Cluster 2 contains 147 genes with reduced expression exclusively in non-ripe fruits that have undergone a long-term cold storage (E3: R
-, C
+), whereas cluster 6 contains 114 genes with increased expression exclusively in these same fruits. Surprisingly, only 4 of the genes, present in cluster 6 correspond to genes identified as being differentially expressed by Ogundiwin et al [
13]. Differentially expressed genes, identified by Ogundiwin et al [
13], appear in clusters 6, 7, 11 and 12. Therefore, we have identified a very large number of novel genes that are associated specifically with the unripe fruits that have undergone long-term cold storage. These genes are associated with a large number of biological processes in multiple regions of the cell. By further analyzing the function of these genes we may be able to better understand the specific modifications that have occurred in the fruit under long-term cold storage that ultimately results in the modification in the ripening process that result in woolly fruits.
1,3 Andrés Tittarelli,2,3 Mariano Latorre,2,3 Juan Saba,3 Rodrigo Caroca,1,3 Jonathan Maldonado,2,3 Veronica Cambiazo,4,5 Reinaldo Campos-Vargas,1,3,6 Mauricio Gonzalez,4 Ariel Orellana,1,3 and Herman Silva
This article has been 
