iMat input / Tutorial example

[Taneli Pusa]

Hi everyone,
I'm trying to figure out the iMat method, its input and the actual preprocessing pipeline.
In Shlomi et al. 2008, as well as the Zur et al. 2010 paper, the authors say about mapping the gene expression data to reactions:
"Specifically, this was done by modifying the Boolean mapping to account for tri-valued expression states, assigning highly, lowly and moderately expressed genes, values of 1, –1 and 0, respectively, and replacing the logical ‘and’ and ‘or’ operators with ‘max’ and ‘min’ expressions, respectively (following ref. 22 [Bilu et al. 2006]).",
emphasis mine.
In the tutorial provided at
https://opencobra.github.io/cobratoolbox/latest/tutorials/tutorialExtraction_transcriptomic.html
and example model, data set, and preprocessed data are given. Looking at this example, taking for example the reaction 'ACKr', it's gene association is
'(b3115 or b2296 or b1849)'
and the mean expression levels for these genes are
b3115  2.5844
b2296  13.1665
b1849  8.7057
However, in the example input for iMat, 'options', the expression level for 'ACKr' is 8.7057.

I don't understand how to arrive at this expression for the reaction, given the gene association and the corresponding levels. The other expression levels in 'options' that I've looked are equally confusing, with some following the logic 'or'=min and some not. So how exactly should the preprocessing be done?

Another question I have is: in the same 'options' example, 'threshold_lb'=-0.2131. How can this be negative? All the expression levels should be positive reals.

Thanks in advance,
Taneli

[Thomas Pfau]

Hi

As for the Expression Data: Seems like the example data is off. We will fix this soonish. Wrt the threshold you are right, this should not be negative.
However this threshold depends a lot on the actual data used and the preference of the user and we only provide a framework of the iMAT algorithm. the createSpecificTissueModel functions requires the threshold to be set by the user for this reason, and we don't want to provide a default here, as this is completely data dependent.
As for the Reaction Expression:
If you use the mapExpressionToReactions function, you will obtain the mapping as described in the paper you mentioned (or -> max; and -> min), and we will update the data in the tutorial accordingly..

Best

Thomas

--

---
You received this message because you are subscribed to the Google Groups "COBRA Toolbox" group.

-- 
Université du Luxembourg
Faculté des Sciences, de la Technologie et de la Communication
Campus Belval, Biotech II 115
6 avenue du Swing
L-4367 Belvaux
Tel: (+352) 46 66 44 5309
Email: thoma...@uni.lu

[Taneli Pusa]

Thank you for your reply Thomas.
What you said makes sense, and indeed the function 'mapExpressionToReactions' produces what one would expect.

However, I do have to point out that this is in fact NOT what is described in the two papers. To quote again, both Shlomi et al. 2008 and Zur et al. 2010 say:
"replacing the logical ‘and’ and ‘or’ operators with ‘max’ and ‘min’ expressions, respectively [...]"
implying and -> max; or -> min. Given that the reference they give, Bilu et al. 2006, does not follow the logic Shlomi et al. describe, and the fact that the contrary makes more sense, I'm assuming this is a typo in both papers.

Kind regards,
Taneli

[Thomas Pfau]

Hi Taneli,

I would assume that to be a typo in the paper as well, because and -> max; or -> min does not might any sense at all.

Best

Thomas

On 03/02/2018 11:36 AM, Taneli Pusa wrote:

Unsubscribe

It appears that you have subscribed to commercial messages from this sender. To stop receiving such messages from this sender, please unsubscribe

To unsubscribe from this group and stop receiving emails from it, send an email to cobra-toolbo...@googlegroups.com.
For more options, visit http://secure-web.cisco.com/1aYrMy1miepqE-OROrn8NLuB9x0f18ptD76wg-RppynicyBCJ63NZRR6gCjRY1zNJhdCjdZhLmDUvnvgqHh9w-kd2tZQEib1Ny_uXC8aVjYR4vBzSKFMkG_w1YFWz-WY38PJmfTvqHmLT2Y6zSSOeLpUGXsZMeZdd0tHC31Mbhige3DHvImm3j_B0ygeosqHD6CJacsXlWDK8S8FOOhoJRb5OMH57NDEPV8SBp946zRo3mmIiwRIJ0E3q5n5hV5_LA6wTaARhNRyuBpj6xCZICzluhp3DZhNEbbuPulyeYhBGoBrmr3RKnrmrYvE6h6NCn02CfK-d9S1oECWyXOuLYYdyUADaO0OU6DsXv3YS4ZxrT6tZsXysf4mLgOmXLbAxcFwbQZotWV14J74pH6Au9w/l34%3Ahttps%3A%2F%2Fgroups.google.com%2Fd%2Foptoute.

-- 
Université du Luxembourg
Faculté des Sciences, de la Technologie et de la Communication
Campus Belval, Biotech II 423
6 avenue du Swing
L-4367 Belvaux
Tel: (+352) 46 66 44 5309
Email: thoma...@uni.lu