5. Acetonitrile (ACN) ULC/MS grade. Caution: Acetonitrile is
toxic and should be handled under the fume hood.
6. Formic acid (HCOOH) ULC/MS grade. Caution: Formic
acid is corrosive and should be handled under the fume hood.
2.3.3 Data Analysis
1. Computer and Software for mass spectral data analysis (see
[26]).
3 Methods
3.1 Experimental
Design
In order to correctly evaluate such large sample sets it is important
to ensure that variables are manipulated under strictly controlled
conditions while taking precise measurements. As such, the precision of an experiment critically depends on the size of the experiment and the homogeneity of experimental materials. In large
genomics experiments, another critical step is to determine the
number of lines and associated biological replicates. This is then
followed by choosing the statistical approach to link genotype with
phenotype. Here we neither focus on the choice of the number of
lines nor the population structure needed to obtain a complete
genotype-to-phenotype matrix in order to identify all possible
QTL (see [27] for more details) but rather on how many biological
replicates are required per line for acceptable statistical analysis and
data normalization (see Note 4).
In QTL analysis the number of replicates profiled will have a
major influence on the reliability and reproducibility of the data and
consequently on the QTL mapping results. Therefore, the ability to
make broad conclusions or identify causal genes using quantitative
studies of metabolic variation is greatly influenced by the fact that
metabolic abundances measured in these studies are highly dependent on the environmental, developmental, and genetic variations
present within the experiment as well as the experimental error.
Based on our own experience, it is recommended to use at least six
independent biological replicates for each line (genotype) and
many more control plants in a completely randomized design to
overcome unavoidable effects associated with variation in microenvironmental factors such as light intensities, temperature, and air
humidity. This should be planed carefully in advance and the population size and time needed for collecting the samples should also
be taken into account in order to ensure that harvesting is carried
out in as rapid and homogeneous a manner as possible.
3.2 Plant Material
and Sampling
1. In the case of introgression lines (ILs) a reasonable number of
biological replicates is six independent plants (see Note 5)
2. Collect several different plant organs (3–5) per biological replicate, pool them, and treat them as a single sample.
Metabolomic Analysis of Natural Variation in Arabidopsis
397
toxic and should be handled under the fume hood.
6. Formic acid (HCOOH) ULC/MS grade. Caution: Formic
acid is corrosive and should be handled under the fume hood.
2.3.3 Data Analysis
1. Computer and Software for mass spectral data analysis (see
[26]).
3 Methods
3.1 Experimental
Design
In order to correctly evaluate such large sample sets it is important
to ensure that variables are manipulated under strictly controlled
conditions while taking precise measurements. As such, the precision of an experiment critically depends on the size of the experiment and the homogeneity of experimental materials. In large
genomics experiments, another critical step is to determine the
number of lines and associated biological replicates. This is then
followed by choosing the statistical approach to link genotype with
phenotype. Here we neither focus on the choice of the number of
lines nor the population structure needed to obtain a complete
genotype-to-phenotype matrix in order to identify all possible
QTL (see [27] for more details) but rather on how many biological
replicates are required per line for acceptable statistical analysis and
data normalization (see Note 4).
In QTL analysis the number of replicates profiled will have a
major influence on the reliability and reproducibility of the data and
consequently on the QTL mapping results. Therefore, the ability to
make broad conclusions or identify causal genes using quantitative
studies of metabolic variation is greatly influenced by the fact that
metabolic abundances measured in these studies are highly dependent on the environmental, developmental, and genetic variations
present within the experiment as well as the experimental error.
Based on our own experience, it is recommended to use at least six
independent biological replicates for each line (genotype) and
many more control plants in a completely randomized design to
overcome unavoidable effects associated with variation in microenvironmental factors such as light intensities, temperature, and air
humidity. This should be planed carefully in advance and the population size and time needed for collecting the samples should also
be taken into account in order to ensure that harvesting is carried
out in as rapid and homogeneous a manner as possible.
3.2 Plant Material
and Sampling
1. In the case of introgression lines (ILs) a reasonable number of
biological replicates is six independent plants (see Note 5)
2. Collect several different plant organs (3–5) per biological replicate, pool them, and treat them as a single sample.
Metabolomic Analysis of Natural Variation in Arabidopsis
397