PathwayCommon gives different results

[Winston Huang]

I tried to analysis the interaction network among a bunch of genes.

However, I got different networking every time...

I need the analysis which is reproducible for paper publication.

Anyone can help?

Thanks

[Igor R]

Hi Winston,

Could you please provide more details, give an example of what you did, what expected and what you got, exactly?

Igor.

--
You received this message because you are subscribed to the Google Groups "pathway-commons-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pathway-commons-...@googlegroups.com.
To post to this group, send email to pathway-co...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pathway-commons-help/e4b701bf-1355-4827-96b0-e227ad56c16d%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Winston Huang]

Hi Igor,

 

Thanks for your quick reply.

 

The genes I was looking for were STAT3, DDIT3, LPL, ANGPT2, SERPINB2, SNAI3 and CA9.

 

When I did the analysis back in 2016/12/29, the results were shown as in attachment.

 

Recently I tried to check interaction again, the result looks awkward (see attachment too).

 

I am expecting STAT3 to have some relationship with DDIT3/LPL/ANGPT2/CA9.

 

Thanks,

 

Winston

 

Sent from Mail for Windows 10

--
You received this message because you are subscribed to a topic in the Google Groups "pathway-commons-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/pathway-commons-help/egoFJ-mn9i4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to pathway-commons-...@googlegroups.com.

To post to this group, send email to pathway-co...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/pathway-commons-help/732CFF07-D746-4C5A-A0B8-EB57DAB85BEA%40gmail.com.

[Igor Rodchenkov]

As I understand, you used Pathway Commons' PCViz app (an older version versus current latest one) - submitted the same list of genes in the input field: STAT3 DDIT3 LPL ANGPT2 SERPINB2 SNAI3 CA9 - and analysed the visualisations (network pictures). Is that correct? (You could also download the query results in several data format for further off-line analysis.) 

So, in fact, you do not get "different networking every time..." (not true), but the problem is that the new version displays not the same network as the older PCViz did when you submit exactly the same gene list, alright?

Yes, current PCViz app and PC pathway data is different from that in our previous releases: 

besides using new pathway commons web services (PC9), unlike previous versions, current PCViz app always executes the "pathsbetween" graph query instead of "neighborhood" when multiple gene names/IDs are submitted, and it also applies more strict filters (e.g., based on the number of co-citations) to prevent huge "hairball' networks or failures/timeouts from happening often.

I think, in your publication, you could simply accurately describe what you did and what you get using which Pathway Commons/PCViz version (or day)...

Or - get the data in SIF format directly from the PC web service using a query like this (which applies the same 3 inference rules that the old PCViz used); I did, and found, e.g.:

DDIT3 controls-state-change-of STAT3
STAT3 controls-expression-of DDIT3

entries are there.

Try also - 'pathsbetween' instead of 'neighborhood' in the above query link.

Or - try submitting gene name/IDs one by one (this makes PCViz app run the 'neighborhood' query).

Hope it helps.

 

On Friday, October 13, 2017 at 1:59:55 AM UTC-4, Winston Huang wrote:

Hi Igor,

 

Thanks for your quick reply.

 

The genes I was looking for were STAT3, DDIT3, LPL, ANGPT2, SERPINB2, SNAI3 and CA9.

 

When I did the analysis back in 2016/12/29, the results were shown as in attachment.

 

Recently I tried to check interaction again, the result looks awkward (see attachment too).

 

I am expecting STAT3 to have some relationship with DDIT3/LPL/ANGPT2/CA9.

 

Thanks,

 

Winston

 

Sent from Mail for Windows 10

 

From: Igor R
Sent: October 12, 2017 11:29 PM
To: pathway-co...@googlegroups.com
Subject: Re: PathwayCommon gives different results

 

Hi Winston,

 

Could you please provide more details, give an example of what you did, what expected and what you got, exactly?

 

Igor.

On Oct 13, 2017, at 12:28 AM, Winston Huang <wyhh...@gmail.com> wrote:

I tried to analysis the interaction network among a bunch of genes.

 

However, I got different networking every time...

I need the analysis which is reproducible for paper publication.

 

Anyone can help?

 

Thanks

--
...