mapExpressionToReactions "-1" output and CORDA2 algorithm

[Moohebat]

Hi all,

I have reconstructed a metabolic network using CORDA2 algorithm which take as input high confidence (HC), medium confidence (MC) and negative confidence (NC) reaction sets. I used mapExpressionToReactions function on my preprocessed gene expression data to get expression values for the reactions in RECON 2.2, then I set aside the reactions with expression value "-1" (as my understanding is that these are the reactions for which no gene expression data is available) and Z-scored the rest of the expression values and divided into HC, MC and NC and fed to CORDA2.

 

I was wondering whether I was wrong with putting aside those reactions with -1 expression value and I cant find anything on what -1 actually means in the output of mapExpressionToReactions in the documentations. I want to know whether I should have included the -1 valued reactions in the NC set. (And please keep in mind that CORDA2 is a flexible algorithm and includes reactions other than HC and MC if they are necessary in the final model)

I would really appreciate your help with this. Thank you very much.

Best,

Moohebat

[Ronan M.T. Fleming]

Hi Moohebat,
in the latest version of the cobra toolbox, mapExpressionToReactions
now returns NaN for any reaction without expression data, e.g., if a
reaction has no genes associated with it, then the corresponding value
will be NaN. I changed it from -1 to NaN, because the latter means no
information, whereas the former could be passed inadvertently into,
e.g., mean() and give an erroneous result, when mean( , 'omitnan')
would be more appropriate.
Regards,
Ronan

> --
>
> ---
> You received this message because you are subscribed to the Google Groups "COBRA Toolbox" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to cobra-toolbo...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/cobra-toolbox/f0a71618-275a-4ac5-9e03-680f7a55e5b9n%40googlegroups.com.



--
--
Mr. Ronan MT Fleming B.V.M.S. Dip. Math. Ph.D.
----------------------------------------------------------------------------
Assistant Professor,
Division of Systems Biomedicine and Pharmacology,
Leiden Academic Centre for Drug Research,
Faculty of Science,
Leiden University.
https://www.universiteitleiden.nl/en/staffmembers/ronan-fleming
&
H2020 Project Coordinator,
Systems Medicine of Mitochondrial Parkinson’s Disease,
http://sysmedpd.eu
&
Senior Lecturer,
School of Medicine,
National University of Ireland, Galway.
----------------------------------------------------------------------------
Peer-reviewed publications: https://goo.gl/FZPG23
Mobile: +353 852 109 806
Skype: ronan.fleming
----------------------------------------------------------------------------
(This message is confidential and may contain privileged information.
It is intended for the named recipient only. If you receive it in
error please notify me and permanently delete the original message and
any copies.)

[Moohebat Pourmajidian]

Dear Ronan,

Thank you very much for your response. I was encountering several errors when using the latest version of cobra toolbox so I had to go ahead with the old version. So if I understood you correctly, -1 does actually mean that there was no expression data for that reaction, and it should be excluded when z-scoring, etc. Thanks again.

Best,

Moohebat

To view this discussion on the web visit https://groups.google.com/d/msgid/cobra-toolbox/CAOivGYtoJpqy%3DCt5o4TK2SRe-iN3w8RjkHkwz%2Be8jEh7Qv7H6Q%40mail.gmail.com.

[Ronan M.T. Fleming]

Hi Moohebat,
without knowing what exactly the "errors" were, we cannot make
"improvements". I do realise that changing the default from -1 to NaN
might create some downstream issues, but those should be possible to
debug with sufficient information on them. However, that is still
better than the risk of proceeding to interpret -1 as an expression
value, when it is supposed to represent no information.
Regards,
Ronan

> To view this discussion on the web visit https://groups.google.com/d/msgid/cobra-toolbox/CAO0rZT7HZnHLE6Qw2a5gM_Yv91atvZ8hA61h%2Bx4wZ6N09Gp5ew%40mail.gmail.com.

[Moohebat Pourmajidian]

Dear Ronan,

Thanks for your explanation. The errors I was encountering weren't about the mapExpressionToReactions function, they occurred while running the CORDA2 algorithm so I had to switch to the old version. I will make sure to report the future issues here to help with the improvements. Thank you very much.

Kind regards,

Moohebat

To view this discussion on the web visit https://groups.google.com/d/msgid/cobra-toolbox/CAOivGYtCr6wgRoU9FHiSMics_9an4U1cgKS5XwZ2g8N67UNS5A%40mail.gmail.com.