Limits...
Tracking the best reference genes for RT-qPCR data normalization in filamentous fungi.

Llanos A, François JM, Parrou JL - BMC Genomics (2015)

Bottom Line: This group included ubcB, sac7, fis1 and sarA genes, as well as TFC1 and UBC6 that were previously validated for their use in S. cerevisiae.We propose a set of 6 genes that can be used as reference genes in RT-qPCR data normalization in any field of fungal biology.However, we recommend that the uniform transcription of these genes is tested by systematic experimental validation and to use the geometric averaging of at least 3 of the best ones.

View Article: PubMed Central - PubMed

Affiliation: Université de Toulouse; INSA, UPS, INP; LISBP, 135 Avenue de Rangueil, F-31077, Toulouse, France. llanos@insa-toulouse.fr.

ABSTRACT

Background: A critical step in the RT-qPCR workflow for studying gene expression is data normalization, one of the strategies being the use of reference genes. This study aimed to identify and validate a selection of reference genes for relative quantification in Talaromyces versatilis, a relevant industrial filamentous fungus. Beyond T. versatilis, this study also aimed to propose reference genes that are applicable more widely for RT-qPCR data normalization in filamentous fungi.

Results: A selection of stable, potential reference genes was carried out in silico from RNA-seq based transcriptomic data obtained from T. versatilis. A dozen functionally unrelated candidate genes were analysed by RT-qPCR assays over more than 30 relevant culture conditions. By using geNorm, we showed that most of these candidate genes had stable transcript levels in most of the conditions, from growth environments to conidial germination. The overall robustness of these genes was explored further by showing that any combination of 3 of them led to minimal normalization bias. To extend the relevance of the study beyond T. versatilis, we challenged their stability together with sixteen other classically used genes such as β-tubulin or actin, in a representative sample of about 100 RNA-seq datasets. These datasets were obtained from 18 phylogenetically distant filamentous fungi exposed to prevalent experimental conditions. Although this wide analysis demonstrated that each of the chosen genes exhibited sporadic up- or down-regulation, their hierarchical clustering allowed the identification of a promising group of 6 genes, which presented weak expression changes and no tendency to up- or down-regulation over the whole set of conditions. This group included ubcB, sac7, fis1 and sarA genes, as well as TFC1 and UBC6 that were previously validated for their use in S. cerevisiae.

Conclusions: We propose a set of 6 genes that can be used as reference genes in RT-qPCR data normalization in any field of fungal biology. However, we recommend that the uniform transcription of these genes is tested by systematic experimental validation and to use the geometric averaging of at least 3 of the best ones. This will minimize the bias in normalization and will support trustworthy biological conclusions.

Show MeSH
Hierarchical ANOVA of the putative reference genes. Three-level, nested ANOVA with ‘genes’ as the first level, ‘culture conditions’ as the second level and ‘biological replicates’ as the third level. As in Figure 2, this ANOVA was carried out using different sets of conditions: ‘C sources’ subset (A); ‘Stress’ subset (B); ‘Germination’ subset (C). Left graph: relative expression values (Log (base 2)) as a function of the different conditions for the different genes, taking as the control conditions the glucose sample (A & B subsets) and the T0 time-point for spore germination (C subset). Two values were used for each condition (i.e. duplicated experiment). Right panel: partitioning of the variance into the three levels (in %).
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4342825&req=5

Fig3: Hierarchical ANOVA of the putative reference genes. Three-level, nested ANOVA with ‘genes’ as the first level, ‘culture conditions’ as the second level and ‘biological replicates’ as the third level. As in Figure 2, this ANOVA was carried out using different sets of conditions: ‘C sources’ subset (A); ‘Stress’ subset (B); ‘Germination’ subset (C). Left graph: relative expression values (Log (base 2)) as a function of the different conditions for the different genes, taking as the control conditions the glucose sample (A & B subsets) and the T0 time-point for spore germination (C subset). Two values were used for each condition (i.e. duplicated experiment). Right panel: partitioning of the variance into the three levels (in %).

Mentions: In the context of conidial germination, the identification of reliable reference genes was first challenged by the difficulty of producing good quality RNA samples. The influence of RNA quality on reproducibility of measured transcript levels was recently reviewed [73,74], highlighting that the process of normalization does not completely resolve the bias of using compromised RNA quality on the final results. In our hands, only the use of the TriZol reagent secured the mRNA quality standard required for reliable RT-qPCR analysis. This technical prerequisite being fulfilled, the analysis of Cq values in this ‘Germination’ subset showed that the M values increased more rapidly than for ‘C-sources’ and ‘Stress’ subsets, indicating higher expression variability of the genes. This was further illustrated by a hierarchical ANOVA of the relative transcript level data (Figure 3), where it was observed that 70 to 80% of the variation for the ‘C-sources’ and ‘Stress’ subsets took place at the level of the biological replicates (Figure 3A & B), supporting the extremely low variation between genes as well as the low influence of conditions on the transcript levels. In contrast, the variation observed between genes strongly increased in the ‘Germination’ subset, to reach about 50% of total variation (Figure 3C), which was particularly emphasized with genes such as R6 and R11 that exhibited a strong bias (higher expression and activation during germination). The genes R10, R2 and R3, which were classified by geNorm as the best reference genes in this specific subset, were used for normalization (see below, NF(R10, R2, R3)) and confirmed that R6 and R11 were induced respectively by 6 and 12-fold, 6 hours after the beginning of the germination process (data not shown).Figure 3


Tracking the best reference genes for RT-qPCR data normalization in filamentous fungi.

Llanos A, François JM, Parrou JL - BMC Genomics (2015)

Hierarchical ANOVA of the putative reference genes. Three-level, nested ANOVA with ‘genes’ as the first level, ‘culture conditions’ as the second level and ‘biological replicates’ as the third level. As in Figure 2, this ANOVA was carried out using different sets of conditions: ‘C sources’ subset (A); ‘Stress’ subset (B); ‘Germination’ subset (C). Left graph: relative expression values (Log (base 2)) as a function of the different conditions for the different genes, taking as the control conditions the glucose sample (A & B subsets) and the T0 time-point for spore germination (C subset). Two values were used for each condition (i.e. duplicated experiment). Right panel: partitioning of the variance into the three levels (in %).
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4342825&req=5

Fig3: Hierarchical ANOVA of the putative reference genes. Three-level, nested ANOVA with ‘genes’ as the first level, ‘culture conditions’ as the second level and ‘biological replicates’ as the third level. As in Figure 2, this ANOVA was carried out using different sets of conditions: ‘C sources’ subset (A); ‘Stress’ subset (B); ‘Germination’ subset (C). Left graph: relative expression values (Log (base 2)) as a function of the different conditions for the different genes, taking as the control conditions the glucose sample (A & B subsets) and the T0 time-point for spore germination (C subset). Two values were used for each condition (i.e. duplicated experiment). Right panel: partitioning of the variance into the three levels (in %).
Mentions: In the context of conidial germination, the identification of reliable reference genes was first challenged by the difficulty of producing good quality RNA samples. The influence of RNA quality on reproducibility of measured transcript levels was recently reviewed [73,74], highlighting that the process of normalization does not completely resolve the bias of using compromised RNA quality on the final results. In our hands, only the use of the TriZol reagent secured the mRNA quality standard required for reliable RT-qPCR analysis. This technical prerequisite being fulfilled, the analysis of Cq values in this ‘Germination’ subset showed that the M values increased more rapidly than for ‘C-sources’ and ‘Stress’ subsets, indicating higher expression variability of the genes. This was further illustrated by a hierarchical ANOVA of the relative transcript level data (Figure 3), where it was observed that 70 to 80% of the variation for the ‘C-sources’ and ‘Stress’ subsets took place at the level of the biological replicates (Figure 3A & B), supporting the extremely low variation between genes as well as the low influence of conditions on the transcript levels. In contrast, the variation observed between genes strongly increased in the ‘Germination’ subset, to reach about 50% of total variation (Figure 3C), which was particularly emphasized with genes such as R6 and R11 that exhibited a strong bias (higher expression and activation during germination). The genes R10, R2 and R3, which were classified by geNorm as the best reference genes in this specific subset, were used for normalization (see below, NF(R10, R2, R3)) and confirmed that R6 and R11 were induced respectively by 6 and 12-fold, 6 hours after the beginning of the germination process (data not shown).Figure 3

Bottom Line: This group included ubcB, sac7, fis1 and sarA genes, as well as TFC1 and UBC6 that were previously validated for their use in S. cerevisiae.We propose a set of 6 genes that can be used as reference genes in RT-qPCR data normalization in any field of fungal biology.However, we recommend that the uniform transcription of these genes is tested by systematic experimental validation and to use the geometric averaging of at least 3 of the best ones.

View Article: PubMed Central - PubMed

Affiliation: Université de Toulouse; INSA, UPS, INP; LISBP, 135 Avenue de Rangueil, F-31077, Toulouse, France. llanos@insa-toulouse.fr.

ABSTRACT

Background: A critical step in the RT-qPCR workflow for studying gene expression is data normalization, one of the strategies being the use of reference genes. This study aimed to identify and validate a selection of reference genes for relative quantification in Talaromyces versatilis, a relevant industrial filamentous fungus. Beyond T. versatilis, this study also aimed to propose reference genes that are applicable more widely for RT-qPCR data normalization in filamentous fungi.

Results: A selection of stable, potential reference genes was carried out in silico from RNA-seq based transcriptomic data obtained from T. versatilis. A dozen functionally unrelated candidate genes were analysed by RT-qPCR assays over more than 30 relevant culture conditions. By using geNorm, we showed that most of these candidate genes had stable transcript levels in most of the conditions, from growth environments to conidial germination. The overall robustness of these genes was explored further by showing that any combination of 3 of them led to minimal normalization bias. To extend the relevance of the study beyond T. versatilis, we challenged their stability together with sixteen other classically used genes such as β-tubulin or actin, in a representative sample of about 100 RNA-seq datasets. These datasets were obtained from 18 phylogenetically distant filamentous fungi exposed to prevalent experimental conditions. Although this wide analysis demonstrated that each of the chosen genes exhibited sporadic up- or down-regulation, their hierarchical clustering allowed the identification of a promising group of 6 genes, which presented weak expression changes and no tendency to up- or down-regulation over the whole set of conditions. This group included ubcB, sac7, fis1 and sarA genes, as well as TFC1 and UBC6 that were previously validated for their use in S. cerevisiae.

Conclusions: We propose a set of 6 genes that can be used as reference genes in RT-qPCR data normalization in any field of fungal biology. However, we recommend that the uniform transcription of these genes is tested by systematic experimental validation and to use the geometric averaging of at least 3 of the best ones. This will minimize the bias in normalization and will support trustworthy biological conclusions.

Show MeSH