geNorm Normalisation

[lmanchon]

--Hi,

In further improvement it would be usefull to add geNorm normalization option to take into account several control genes. Thank you --

Laurent --

[Stefan-Hinz]

Are you still planning to add this feature to the package? I would agree that this would be a useful option.

Stefan

[MahShaaban]

Hi @stefanCoH Yes. I am planning to add the feature to the package. In fact, I already add code for two functions pcr_nf and pcr_genorm to calculate a normalization factor from multiple reference genes and use the normalization factor to calculate relative expression for a gene of interest. Right now, this feature is in a devel branch multiref and not included in the CRAN release. To include it in the release I need to do the following:

• Find a dataset to test and bench mark the calculation
• Document the geNorm method in the vignette

If you have any suggestion for either task, your help would be very appritiated.

Thanks

[lmanchon]

--Hi Mahmoud,

i have a dataset i can send you to test the calculation, this dataset was used by Vandesompele et al. to develop their algorithm and to determine the most stable reference (housekeeping) genes from a set of tested candidate reference genes in a given sample panel. I send you the dataset tomorrow from my work place. see you, Laurent --

[MahShaaban]

Thanks a lot @lmanchon. I'd be very grateful.

[lmanchon]

--Hi, you can dowload the file directly from this link: https://static-content.springer.com/esm/art%3A10.1186%2Fgb-2002-3-7-research0034/MediaObjects/13059_2001_453_MOESM1_ESM.txt related paper: https://doi.org/10.1186/gb-2002-3-7-research0034

Laurent --

[lmanchon]

geNorm is well documented here: https://genorm.cmgg.be/

[MahShaaban]

Thanks @lmanchon. Actually, I am tried to use this dataset before but the issue was, it only has the relative expression values, not the Ct values. Also it has data for several reference genes but no gene of interest.

What I am looking for is a dataset of multiple reference genes and at least one gene of interest. Preferably where the relative expression was calculated using geNorm and published.

[lmanchon]

--Hi, right. Maybe you can download raw Cq from this page: https://doi.org/10.1371/journal.pone.0122515 under table S1.

[Stefan-Hinz]

maybe this dataset can help? extracted from PMCID: PMC4794502

[MahShaaban]

Thanks @lmanchon and @stefanCoH. This is very helpful.

[digs1998]

The approach towards doing a PCR is really appreciable. My query is, is there any way these can be done in python language.

[lmanchon]

--Yes, you can do it, try with this tool: https://github.com/zqfang/QPCR good luck

[MahShaaban]

I tried to provide the same implementation in a python package. Here, https://github.com/MahShaaban/pycr

[digs1998]

Thank you for your response

On Mon 30 Mar, 2020, 06:43 Mahmoud Shaaban, notifications@github.com wrote:

I tried to provide the same implementation in a python package. Here, https://github.com/MahShaaban/pycr

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/MahShaaban/pcr/issues/4#issuecomment-605733850, or unsubscribe https://github.com/notifications/unsubscribe-auth/AMPSKXELDJEULQZUZWOKBNTRJ7W5HANCNFSM4F3QKBIA .

[janstrauss1]

Hi @MahShaaban,

I really like pcr but would also be very interested to apply GeNorm normalization accounting for multiple reference genes. Did you make any progress to implement this feature in a future CRAN release?

Also, I wondered if there's a chance to implement some additional algorithms (e.g. GeNorm, NormFinder, BestKeeper...) to select most suitable reference genes?

Many thanks,

Jan

[MahShaaban]

Hi @janstrauss1 I tried to implement GeNorm and algorithms for selection suitable references, and made some progress (multiref). It is not ready however for a CRAN release. The main issue here is finding proper test data. Meaning, I need a dataset where these analyses were performed and the results are known to compare to my implementation. If you are familiar with such dataset, please, let me know.

[janstrauss1]

Hi @MahShaaban,

Many thanks for your effort to implement different algorithms for reference gene selection!

I'm currently working on my own qPCR dataset (see attached ct_data file with anonymized gene names) that is likely to get published as part of a paper soon. You're welcome to use it as a (real) test data set!

The general experimental design is

• 2 x 96-well plates
• 18 time-series samples in total (9 treatment and 9 control, each consisting of triplicates at t1, t2, t3)
• 4 target genes
• 5 candidate reference genes

I've used RefFinder available at https://www.heartcure.com.au/reffinder/ to evaluate the stability of the different candidate reference genes and the geNorm algorithm implement in RefFinder selects ref_gene1 and ref_gene4 in my dataset as the most stable genes while other algorithms seem to select other reference genes.

Maybe this data set helps to get your implementation of different normalization algorithms ready for a CRAN release?

ct_data.txt

[MahShaaban]

Thanks @janstrauss1 I really appreciate you sharing this dataset and congrates for the paper.