[Hobbe]

I am having problems using SSPACE basic with my 454 paired-end data, and was hoping to get some help here. SSPACE runs fine using my Illumina PE data, but my 454-data has much longer insert-sizes (3-5 kb), and I think they really could make difference.

My problem is that SSPACE reads all the 454-pairs in, removes quite a lot of them as the include Ns, and then maps 0 of them. The report is below. It was difficult to get the reads in a format that SSPACE accepts, and I guess that the problem lies in the fastq-files. Some (very few) reads are too long (over 1024 bases), and bowtie complains about these. Would this crash the whole run? I know that bowtie is not the best choice for longer reads, but I thought it would still manage to map some reads? Is SSPACE premium the answer?

Any/all help would be much appreciated,
Henrik

READING READS Lib454:
------------------------------------------------------------
Total inserted pairs = 1217215
Number of pairs containing N's = 1066178
Remaining pairs = 151037
------------------------------------------------------------
...

LIBRARY Lib454 STATS:
################################################################################

MAPPING READS TO CONTIGS:
------------------------------------------------------------
Number of single reads found on contigs = 0
Number of pairs used for pairing contigs / total pairs = 0 / 0
------------------------------------------------------------

READ PAIRS STATS:
Assembled pairs: 0 (0 sequences)
Satisfied in distance/logic within contigs (i.e. -> <-, distance on target: 3709 +/-927.25): 0
Unsatisfied in distance within contigs (i.e. distance out-of-bounds): 0
Unsatisfied pairing logic within contigs (i.e. illogical pairing ->->, <-<- or <-->): 0
---
Satisfied in distance/logic within a given contig pair (pre-scaffold): 0
Unsatisfied in distance within a given contig pair (i.e. calculated distances out-of-bounds): 0
---
Total satisfied: 0 unsatisfied: 0


Estimated insert size statistics (based on 0 pairs):
Mean insert size = 0
Median insert size = 0
REPEATS:
Number of repeated edges = 0
------------------------------------------------------------

################################################################################

[gaffa]

yes, this is possible. The file should be in a TAB delimited format like:

<contig1> <startpos_on_contig1> <endpos_on_contig1> <contig2> <startpos_on_contig2> <endpos_on_contig2>

E.g.
contig1 100 150 contig1 350 300
contig1 4000 4050 contig2 110 60

There is a script in the 'tools' directory of the package to convert SAM/BAM to a tab format.

Regards,
Boetsie

Ok, great! Thanks.

On a slightly related note, how well do you think SSPACE would deal with scaffolding information from other sources than paired/mate-reads, such as e.g. physical/genetic linkage data (supplied then in the above file format)? Some scaffolders (notably Bambus) claim to be able to work with essentially any kind of link information between contigs - could the same be said of SSPACE?

[boetsie]

I am having problems using SSPACE basic with my 454 paired-end data, and was hoping to get some help here. SSPACE runs fine using my Illumina PE data, but my 454-data has much longer insert-sizes (3-5 kb), and I think they really could make difference.

My problem is that SSPACE reads all the 454-pairs in, removes quite a lot of them as the include Ns, and then maps 0 of them. The report is below. It was difficult to get the reads in a format that SSPACE accepts, and I guess that the problem lies in the fastq-files. Some (very few) reads are too long (over 1024 bases), and bowtie complains about these. Would this crash the whole run? I know that bowtie is not the best choice for longer reads, but I thought it would still manage to map some reads? Is SSPACE premium the answer?

SSPACE basic does not handle 454 reads well, simply because the reads are too long for bowtie to align (up to 1024 bases). Also, bowtie can handle only up to two mismatches. In SSPACE premium I've added the BWA-SW aligner to deal with larger reads. Otherwise, you can align the reads yourself with BWA-SW and try to convert the resulting .SAM file to a .tab file (see post above).

Boetsie

[boetsie]

Ok, great! Thanks.

On a slightly related note, how well do you think SSPACE would deal with scaffolding information from other sources than paired/mate-reads, such as e.g. physical/genetic linkage data (supplied then in the above file format)? Some scaffolders (notably Bambus) claim to be able to work with essentially any kind of link information between contigs - could the same be said of SSPACE?

I've not tested it myself, but I think any linking information is suited. I would suggest to give it a try

[is41985]

Hi,

to save me the hassle of going through the code, I have a short question regarding insert sizes.
When scaffolding, does SSPACE use the user specified insert size (from the library.txt file), or the estimated insert size (that is reported in the summary file)?
It is important, since in my case these two seem to differ, and I need the real (user-specified) value to be used.


Thank you,
Ivan.

[boetsie]

Hi,

to save me the hassle of going through the code, I have a short question regarding insert sizes.
When scaffolding, does SSPACE use the user specified insert size (from the library.txt file), or the estimated insert size (that is reported in the summary file)?
It is important, since in my case these two seem to differ, and I need the real (user-specified) value to be used.


Thank you,
Ivan.

Hi Ivan,

Ha, don't do that, posting it here will save you some time SSPACE uses the user specified insert-size. The insert size in the report is just some extra information.

Regards,
Boetsie

[is41985]

Hi Ivan,

Ha, don't do that, posting it here will save you some time SSPACE uses the user specified insert-size. The insert size in the report is just some extra information.

Regards,
Boetsie

Indeed it did
Thank you for the quick reply!


Best regards,
Ivan.

[Hiroki]

Hi boetsie,

I am trying to make scaffold from contigs from 454 data with Illumina mate pair 75-bp reads with different sizes.

There are 502k contigs and only 1000 or so decreased by scaffolding. This was much less than I expected.

My current parameters are as follows:
-x = 0
-z = 0
-k = 5
-a = 0.7
-n = 15
-T = 1
-p = 0

Which option should I change??

Of course I know there is a possibility that my mate pair libraries are crap. But I just want to try different settings.

Sorry for a newbie question.

Thank you so much

[boetsie]

It is hard to tell, but probably you have to change something in your library file. To give you further advice I both need the output summaryfile and input library file for this. Could you maybe send them through e-mail ([email protected])?

Boetsie

Hi boetsie,

I am trying to make scaffold from contigs from 454 data with Illumina mate pair 75-bp reads with different sizes.

There are 502k contigs and only 1000 or so decreased by scaffolding. This was much less than I expected.

My current parameters are as follows:
-x = 0
-z = 0
-k = 5
-a = 0.7
-n = 15
-T = 1
-p = 0

Which option should I change??

Of course I know there is a possibility that my mate pair libraries are crap. But I just want to try different settings.

Sorry for a newbie question.

Thank you so much

[Hiroki]

Thanks for the response.
I will PM you soon.

Hiroki

[Hiroki]

Hi Boetsie,

I emailed you some files a couple days ago.
Did you receive them?
Let me know if anything wrong.

Sorry if I'm rushing you...

Thanks.

Hiroki

[gaffa]

Is there any easy way to find out for a given SSPACE scaffold what are the names of the pre-SSPACE contigs/scaffolds that are included in it?

[boetsie]

Is there any easy way to find out for a given SSPACE scaffold what are the names of the pre-SSPACE contigs/scaffolds that are included in it?

Hi Gaffa,

have a look in the 'intermediate_results' folder, there is a file ending with *formattedcontigs.fasta. Here, the original headers are named after 'seed'.

Boetsie

[rpdias]

Greetings,
I'm quite new to SSPACE. I'm using it in a Ubuntu 64bits VM machine with perl and python.
I'm tying to use the example data to test SSPACE, but I have always the same error, it seems that can't do bowtie indexing. The same happens with original Illumina data.

Anyone had the same problem or can help me on this?

Thanks in advance for any help!

Cheers,
Ricardo


SSPACE Log
perl /media/NGStorage/NGS/SW/SSPACE-BASIC-2.0_linux-x86_64/SSPACE_Basic_v2.0.pl -l libraries.txt -s contigs_abyss.fasta -k 5 -a 0.7 -x 0 -b ecoli_scaffolds_no_extension

Legacy library getopts.pl will be removed from the Perl core distribution in the next major release. Please install the separate libperl4-corelibs-perl package. It is being used at /media/NGStorage/NGS/SW/SSPACE-BASIC-2.0_linux-x86_64/SSPACE_Basic_v2.0.pl, line 87.
Your inserted inputs on [SSPACE_Basic_v2.0_linux] at Tue Aug 21 10:09:36 2012:
Required inputs:
-l = libraries.txt
-s = contigs_abyss.fasta
-b = ecoli_scaffolds_no_extension

Optional inputs:
-x = 0
-z = 0
-k = 5
-a = 0.7
-n = 15
-T = 1
-p = 0


=>Tue Aug 21 10:09:36 2012: Reading, filtering and converting input sequences of library file initiated
Reading read-pairs lib1.1 @ 10000000

------------------------------------------------------------

=>Tue Aug 21 10:10:11 2012: Storing contigs to format for scaffolding

LIBRARY lib1
------------------------------------------------------------

=>Tue Aug 21 10:10:12 2012: Reading contig file

=>Tue Aug 21 10:10:12 2012: Building Bowtie index for contigs

Bowtie-build error; -1 at /media/NGStorage/NGS/SW/SSPACE-BASIC-2.0_linux-x86_64/bin/PairingAndScaffolding.pl line 829.
**************************************************

Process failed on Tue Aug 21 10:10:12 2012

[boetsie]

Hi Ricardo,

look at this post where colindaven suggests how to fix the problem;

http://seqanswers.com/forums/showpost.php?p=50912&postcount=114

Simply chmod a+x all directories of SSPACE.

Regards,
Boetsie