[
./productpagepag.html]
[
./indexpag.html]
[
./support_pagepag.html]
[
./contactpagepag.html]
Copyright © 2005 - MRC-Holland
[
./indexpag.html]
Home
-
[
./productpagepag.html]
Products
-
Support
-
[
./article_pagepag.html]
Articles
-
[
./contactpagepag.html]
contact
MLPA analysis / Normalisation
[
./coffalyser_pagepag.html]
.
[
./order_infopag.html]
[
http://www.mlpa.com/pages/support_searchpag.html]
About MLPA Analysis / Normalisation
Analysis of a limited number of samples can easily be done by visual examination of the capillary electrophoresis peak profiles. A two colour, on screen, overlay of sample and control reaction can be accomplished by e.g. Genotyper software. For the analysis of large numbers of samples, however, exportation of peak areas and using relaible normalisation methods will be necessary.
Normalisation of MLPA data is essential because variations in experimental conditions may lead to quantitative differences of measured values between samples. After normalisation, however, the measured data should reveal the biological differences in copy numbers -- not possible differences in the experimental process. Normalisation thus tries to minimise the amount of systematic variation in the data. By minimising the amount of non-biological variation, it will be possible to focus on real biological changes during data analysis. Systematic variation (variation introduced during extraction, labelling, pipetting, PCR, etc) will affect different experiments to different extents. Normalisation may be applied to the data of two experiments, in most cases a normal control vs (patient) sample. After the normalisation of the two runs, you should be able to compare (for instance) the experimental data to your reference sample. After normalisation of the fluorescence intensities between experiments, you can compare the data from one experiment with another.
Many different methods exist for the analysis of MLPA data, and would be impossible to single out one strategy that works best for all mixes. Every mix should therefore be treated separately. The main assumption of normalisation is a functional coherence between a true biological difference and the corresponding measured values, which should result in an unequivocal function. The second assumption is the presence of a property within the samples which is unchanged. This could be the total peak area (globalisation methods) or the unchanged values of several control genes. The identification of this value that is not affected by experimental conditions and, thus, show a ratio equal to 1 between the reference (control) and the experimental sample. To compare measurements from MLPA experiments, peak area’s (peak height can also be used but are more vulnerable to sloping) can be normalised using our control genes or global normalisation methods based on mean or median values.
MLPA Globalisation Normalisation
MLPA global normalisation is probably the most easiest way of normalizing however this method should only be applied for single gene deletion and not when deletions of large numbers of probes are expected.
We export the size and peak areas obtained to an excel file. Aspecific amplification products, primer dimers and the peaks obtained from the MLPA control mix can usually be removed by their low peak area or small lengths. When only the peak areas of the expected MLPA products are left, all peak areas are normalised by dividing the peak area by the combined peak area of all peaks in that lane.
These normalised peak areas are compared to the results of a control run (usually DNA of a normal Individual) or the average results obtained on all samples. First the normalised peak areas are divided by the average normalised peak area of that probe amplification product of all samples.
Raw data of samples that give many aberrations are checked. Abnormal low peaks of both the MLPA products as well as the size standard may indicate failure of injection in the capillary. Low MLPA products peaks but high residual primer peak may indicate failure of the PCR reaction. High MLPA control peaks (64, 70, 76, 82 bp) but low MLPA product peaks indicate insufficient sample DNA or failure of the ligation reaction.
Results of samples with non reliable results are removed. Again, for each sample, the normalised peak areas are divided by the average normalised peak area of that probe amplification product of all samples. These results are visualised using Excel graphics. A deletion of one copy of a probe target sequence will usually be apparent by a reduction in relative peak area for that probe amplification product of 35-55 %. A gain in copy number from two to three copies / diploid genome will usually be apparent by an increase in relative peak area between 30 and 55 %. Standard deviation of peak areas should be below 10 % for all probes.
MLPA global normalisation can be set in both the Coffalyser and all analysis sheets offered above.
MLPA Normalisation with a Gene-subset
Most MLPA probe mixes contain control probes. These probes are placed in chromosomal regions which are assumed to stay normal in both patient and control and therefore provide a unchanged values which can be used for normalisation. By dividing every peak area result by the area result of a control gene in that same run a normalised number is made which can be used to compare control and sample. The result of this comparison is a ratio balance between control and sample.
Since even control genes may change (depending on the type of samples used), multiple control genes are used, all resulting in a calculated ratio. In the Coffalyser Macro and analysis sheets a median is being taken of all these calculated ratio’s. By taking a median any extremely changed values in a control gene will not affect the calculated end ratio.
MLPA Normalisation using all genes
Information about the used control genes in the MLPA kits can be found in either the mix description or can be acquired by emailing to info@mlpa.com. Control genes should fit your experimental design, in other words, the control genes used in the MLPA kits should be normal in your samples. This is important for you normalization. If knowledge about the chromosomal state of your samples is present, it may be usefull to check whether or not the control genes are in “quiet” chromosomal areas.
If your control genes do not behave as expected you may want to choose for another kind of normalisation. Global normalisation or a normalisation where you use all probes (genes) for normalisation. By doing so, every probe will be assusmed to be a control probes, and will therefore provide us with a ratio. If next a median is being taken from this subset of ratio’s a final ratio is calculated, which should be a correct ratio if your samples do not contain too much abberations. In other words, using all probes as control probes will only be a succesfull strategy if less than 50% of your probes are expected to contain abberations.
MLPA normalisation using all genes can be set in both the Coffalyser and all analysis sheets offered above.
User Defined MLPA normalisation
Not every MLPA mix contains Control genes, and not always will a Global normalization or a normalisation using all probes be succesfull. Therefore all MLPA mixes have the possibility to creat a User Defined normalisation. In this case every probes will be normalized by a certain subset of chosen probes in the same mix. For example our prenatal kit for aneuploidy doesnt contain any control probes, and since a large number of probes are either deleted or gained in the expected patient samples a global normalisation will not be appropriate either. We can therefore choose to create a subset of probes separtely for every probe. In the case of an aneuploidy kit is it possible to normalize every probe on chromosome 21 with the probes on chromosome 18 and 13, and normalize probes on chromosome 18 with the probes on chromosome 21 and 13 etc. By choosing for this type of normalization, we say in other words that for instance in case of a triploidy 21, probes on chromosome 18 and 13 are expected to remain diploid.
[./support_mlpa_infopag.html]
[./support_mlpa_protocolspag.html]
[./support_mlpa_analysispag.html]
[./support_mlpa_analysis_overigpag.html]
[./support_ms_mlpapag.html]
[http://www.mlpa.com/coffalyser/]
[./support_faqpag.html]
[./support_desing_synthetic_probespag.html]
[./support_mlpa_analysis_normalisationpag.html]
[./article_pagepag.html]
[Web Creator]
[LMSOFT]