Calculation of pan- and core-genome

Stegger · 02-08-2012, 11:52 PM

Hi,
I was hoping someone here could point me in the direction of a good tool for calculation of pan and core genomes in prokaryotes! I am looking for one or several tools/scripts that does a number of things:
I have a bucket full of bacterial genome data (in contigs mostly) of the same species and would like, based on various gropings of these, to determine initially overall pan and core genomes of the isolates.

Besides getting just the number of genes in each group, it would also be very beneficial to some sort if genes list output for further analysis.

Finally, I would like to see what difference there is between the calculated core/pan genome in 1 group compared to another defined set of isolates in another group - again not only a number of genes but an actual list of genes or gene sequences.

The contigs have not been analysed for CDSs or annotated in any way, but this I can do in another pipeline prior to the pan core calculation if needed.

Thanks!!!

Zam · 02-09-2012, 12:02 AM

Can I just clarify - do you have a bunch of reads which are labelled with which isolate of the same species they come from, and on the basis of that you want to pull out the pan (everything) and core (shared) genomes?

Stegger · 02-09-2012, 12:13 AM

Hi Zam,
I have assembled the reads into contigs, and they have names_contigID to them indicating the species and specific isolate they come from. And yes, it from them that I would like to extract the information.

Zam · 02-09-2012, 12:25 AM

Well then, one approach is to assemble a "multicoloured" graph of your data (one colour per isolate), and then dump contigs with information about how many isolates share each contig. Then you can split things however you like - pull out the contigs that everyone shares, 95% share, etc. Software for this is here:
http://cortexassembler.sourceforge.net/
and the paper contain an example of something similar:
http://dx.doi.org/10.1038/ng.1028

>Finally, I would like to see what difference there is between the calculated core/pan >genome in 1 group compared to another defined set of isolates in another group - >again not only a number of genes but an actual list of genes or gene sequenc

You can do any comparisons you like between any subsets you like in this manner. Feel free to contact me directly (zam AT well.ox.ac.uk)

Adjuvant · 02-09-2012, 06:37 AM

Good references for how to do the calculations are Kittichotriat W et al, PLoS ONE July 2011 and Tettelin H. et al PNAS 2005 102:13950-13955 if you want to try doing the analysis or scripting out your own tools. There's also Pan Seq that you can try, but I haven't really been able to get it to work all that well for my purposes.

Stegger · 02-10-2012, 04:06 AM

Thanks both of you!!

And Zam, I may take you up on that offer. And congratulations on that paper.

koadman · 02-11-2012, 05:12 AM

At the risk of being accused of shameless self-promotion, I will point out that this is something that Mauve and specifically progressiveMauve has supported for years. Have a look a the .backbone file output (documentation here).

Zam · 02-11-2012, 05:24 AM

Koadman - Good for you! (I'm certainly in no position to criticise self-promotion)
Stegger - thanks!

Stegger · 02-11-2012, 08:53 PM

Please self-promote all you can, that just allow me to come back with potential questions to the right people

shashankgupta · 04-09-2015, 09:28 PM

Quote:

Originally Posted by Zam

Well then, one approach is to assemble a "multicoloured" graph of your data (one colour per isolate), and then dump contigs with information about how many isolates share each contig. Then you can split things however you like - pull out the contigs that everyone shares, 95% share, etc. Software for this is here:
http://cortexassembler.sourceforge.net/
and the paper contain an example of something similar:
http://dx.doi.org/10.1038/ng.1028

>Finally, I would like to see what difference there is between the calculated core/pan >genome in 1 group compared to another defined set of isolates in another group - >again not only a number of genes but an actual list of genes or gene sequenc

You can do any comparisons you like between any subsets you like in this manner. Feel free to contact me directly (zam AT well.ox.ac.uk)

I am also trying to do analysis for PAN/CORE genome, but the above mentioned software is for someone who have good hands in linux based system.

Is there a simple way where non-bioinformatician can do this kind of analysis ?

Cheers !
Shashank

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
qPCR of H4 pan ac ChIP	emilyindallas	SOLiD	0	12-21-2011 07:57 AM
coverage calculation	arvi8689	Illumina/Solexa	7	11-11-2011 02:53 PM
3.4GHz Quad-Core Intel Core i7 versus 3.1GHz Quad-Core Intel Core i5	brachysclereid	Bioinformatics	2	05-03-2011 05:31 PM
depth calculation	sheilal	Bioinformatics	5	10-04-2009 11:20 AM
Genome-wide detection and analysis of hippocampus core promoters using DeepCAGE.	doxologist	Literature Watch	0	01-12-2009 09:33 AM

02-08-2012, 11:52 PM	#1
Stegger Member Location: Copenhagen Join Date: Nov 2008 Posts: 21	Calculation of pan- and core-genome Hi, I was hoping someone here could point me in the direction of a good tool for calculation of pan and core genomes in prokaryotes! I am looking for one or several tools/scripts that does a number of things: I have a bucket full of bacterial genome data (in contigs mostly) of the same species and would like, based on various gropings of these, to determine initially overall pan and core genomes of the isolates. Besides getting just the number of genes in each group, it would also be very beneficial to some sort if genes list output for further analysis. Finally, I would like to see what difference there is between the calculated core/pan genome in 1 group compared to another defined set of isolates in another group - again not only a number of genes but an actual list of genes or gene sequences. The contigs have not been analysed for CDSs or annotated in any way, but this I can do in another pipeline prior to the pan core calculation if needed. Thanks!!!

02-09-2012, 12:02 AM	#2
Zam Member Location: Oxford Join Date: Apr 2010 Posts: 51	Can I just clarify - do you have a bunch of reads which are labelled with which isolate of the same species they come from, and on the basis of that you want to pull out the pan (everything) and core (shared) genomes?

02-09-2012, 12:13 AM	#3
Stegger Member Location: Copenhagen Join Date: Nov 2008 Posts: 21	Hi Zam, I have assembled the reads into contigs, and they have names_contigID to them indicating the species and specific isolate they come from. And yes, it from them that I would like to extract the information.

02-09-2012, 12:25 AM	#4
Zam Member Location: Oxford Join Date: Apr 2010 Posts: 51	Well then, one approach is to assemble a "multicoloured" graph of your data (one colour per isolate), and then dump contigs with information about how many isolates share each contig. Then you can split things however you like - pull out the contigs that everyone shares, 95% share, etc. Software for this is here: http://cortexassembler.sourceforge.net/ and the paper contain an example of something similar: http://dx.doi.org/10.1038/ng.1028 >Finally, I would like to see what difference there is between the calculated core/pan >genome in 1 group compared to another defined set of isolates in another group - >again not only a number of genes but an actual list of genes or gene sequenc You can do any comparisons you like between any subsets you like in this manner. Feel free to contact me directly (zam AT well.ox.ac.uk)

02-09-2012, 06:37 AM	#5
Adjuvant Member Location: Chicago, IL Join Date: Sep 2010 Posts: 13	Good references for how to do the calculations are Kittichotriat W et al, PLoS ONE July 2011 and Tettelin H. et al PNAS 2005 102:13950-13955 if you want to try doing the analysis or scripting out your own tools. There's also Pan Seq that you can try, but I haven't really been able to get it to work all that well for my purposes.

02-10-2012, 04:06 AM	#6
Stegger Member Location: Copenhagen Join Date: Nov 2008 Posts: 21	Thanks both of you!! And Zam, I may take you up on that offer. And congratulations on that paper.

02-11-2012, 05:12 AM	#7
koadman Member Location: Sydney, Australia Join Date: May 2010 Posts: 65	At the risk of being accused of shameless self-promotion, I will point out that this is something that Mauve and specifically progressiveMauve has supported for years. Have a look a the .backbone file output (documentation here).

02-11-2012, 05:24 AM	#8
Zam Member Location: Oxford Join Date: Apr 2010 Posts: 51	Koadman - Good for you! (I'm certainly in no position to criticise self-promotion) Stegger - thanks!

02-11-2012, 08:53 PM	#9
Stegger Member Location: Copenhagen Join Date: Nov 2008 Posts: 21	Please self-promote all you can, that just allow me to come back with potential questions to the right people