SOAP denovo assembly results

bioinfosm · 09-08-2009, 08:46 AM

Hi,

I was able to assemble more than 100 million reads to contigs using SOAP. That was cool, as all other tools ran into memory issues..

However, I wish to understand some of the properties of contigs, like
- how many reads were actually used in the assembly
- what kind of depth of coverage do the contigs have from overlapping reads
- other properties to determine how confident I could be of the contigs

any pointers... and any help with extracting info from the soap assembly results (except the contig sequences I already have)

henry · 09-09-2009, 04:46 AM

Quote:

Originally Posted by bioinfosm

Hi,

I was able to assemble more than 100 million reads to contigs using SOAP. That was cool, as all other tools ran into memory issues..

However, I wish to understand some of the properties of contigs, like
- how many reads were actually used in the assembly
- what kind of depth of coverage do the contigs have from overlapping reads
- other properties to determine how confident I could be of the contigs

any pointers... and any help with extracting info from the soap assembly results (except the contig sequences I already have)

This is a good question. I also wanna know. could anyone kindly help us?
Btw, how much time did it take to denovo assemble more than 100 million reads into contigs?

Best

Jing

lcollado · 09-09-2009, 06:17 AM

How much RAM did you need with SOAP? I'm curious ^^

bioinfosm · 09-09-2009, 09:44 AM

Less than 100Gb RAM, I did not track it (as long as it was not crashing, I was happy)
Time, it took less than a day to get done with its various steps. A lot of contigs are length 24 and definitely not useful.
>1 length 24 cvg_0_tip_0
AAAAAAAAAAAAAAAAAAAAAAAA
>3 length 24 cvg_0_tip_0
AAAAAAAAAAAAAAAAAAAAAAAC
...
>347 length 65 cvg_10_tip_0
TTCAGTAATAACGGCAGACTAATCACCTCAGAAAACACAAAGCACAAGCTTGTGCTTGTCACTTC

Looking for some documentation to understand this better...

node · 09-11-2009, 02:15 AM

A better way to ask for help about SOAPdenovo , is join to SOAP's mailing list !

And here is the SOAP site: http://soap.genomics.org.cn . You can submit your email on the home page .

PS: SOAP use Google group as it's mailing list .

rubi · 05-02-2011, 08:50 AM

Hi,

Can anyone refer me to good documentation on quality assessment of soap denovo assembly (e.g., n50 values, how to compute coverage, and what the output files mean, etc...)

Thanks

y.divyatej@gmail.com · 10-03-2011, 03:00 PM

I'm curious if there is a pipeline available for Soap De novo assembly. Is there a requirement for the number of genomes required for Denovo assembly?

narain · 10-13-2011, 02:14 AM

Dear Members

I am doing de-novo assembly of human genome from fastq data files. I get contigs as well as scaffolds from tools that I use. I know that scaffolds are a combination of contigs with estimated gaps in between them. Does this mean that downstream analysis when comparing it to another genome such as the reference should be done with contigs more reliably than with scaffolds ?

Aby

darren.cullerne · 02-01-2012, 03:39 PM

bioinfosm:

Quote:

However, I wish to understand some of the properties of contigs, like
- how many reads were actually used in the assembly
- what kind of depth of coverage do the contigs have from overlapping reads
- other properties to determine how confident I could be of the contigs

The way our "lab" (aka office) has looked at the number of reads used in an assembly is to take the raw reads and back align them against the newly created denovo contigs. It should give a pretty good indication. We use Kanga for our back alignments. Damned quick and efficient:
http://code.google.com/p/biokanga/

sagarutturkar · 06-13-2012, 08:22 AM

I am confused about contig numbers reported by Soap.I was trying to
run SoapDenovo for small microbial genomes (Size varies from 5-10MB).

However, the number of contigs reported in .contig files is very high
(always in thousands) whereas other assemblers giving me contigs less
than 500. But Soap-Scaffolding output was better than other
assemblers.

I am not sure, if I am looking at some intermediate contig file OR
contig number is always high in Soap? From my experience contig and
Scaffolds numbers differ by few 100s only (at least for small microbial genomes). There is no drastic change, but in Soap contigs were in 2000-3000 range while scaffolds were in 200-500 range. Please explain.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
running soap denovo	sweet_dna_girl	Bioinformatics	6	07-30-2013 07:30 AM
Segmentation fault (core dumped) at contig step during SOAP denovo assembly	tangzhonghui	Bioinformatics	1	10-09-2012 06:32 PM
Rank In Soap denovo	sivasubramani	Introductions	1	08-31-2011 11:05 PM
SOAP denovo output to AFG format	Autotroph	Bioinformatics	0	02-21-2011 02:00 PM
Need help on soap denovo	sundar	De novo discovery	4	11-29-2010 04:03 AM

09-08-2009, 08:46 AM	#1
bioinfosm Senior Member Location: USA Join Date: Jan 2008 Posts: 482	SOAP denovo assembly results Hi, I was able to assemble more than 100 million reads to contigs using SOAP. That was cool, as all other tools ran into memory issues.. However, I wish to understand some of the properties of contigs, like - how many reads were actually used in the assembly - what kind of depth of coverage do the contigs have from overlapping reads - other properties to determine how confident I could be of the contigs any pointers... and any help with extracting info from the soap assembly results (except the contig sequences I already have) __________________ -- bioinfosm

09-09-2009, 06:17 AM	#3
lcollado Member Location: Baltimore, MD Join Date: Jun 2009 Posts: 65	How much RAM did you need with SOAP? I'm curious ^^ __________________ L. Collado Torres, Ph.D. student in Biostatistics.

09-09-2009, 09:44 AM	#4
bioinfosm Senior Member Location: USA Join Date: Jan 2008 Posts: 482	Less than 100Gb RAM, I did not track it (as long as it was not crashing, I was happy) Time, it took less than a day to get done with its various steps. A lot of contigs are length 24 and definitely not useful. >1 length 24 cvg_0_tip_0 AAAAAAAAAAAAAAAAAAAAAAAA >3 length 24 cvg_0_tip_0 AAAAAAAAAAAAAAAAAAAAAAAC ... >347 length 65 cvg_10_tip_0 TTCAGTAATAACGGCAGACTAATCACCTCAGAAAACACAAAGCACAAGCTTGTGCTTGTCACTTC Looking for some documentation to understand this better... __________________ -- bioinfosm

09-11-2009, 02:15 AM	#5
node Member Location: Shenzhen,China Join Date: Sep 2009 Posts: 10	Hi A better way to ask for help about SOAPdenovo , is join to SOAP's mailing list ! And here is the SOAP site: http://soap.genomics.org.cn . You can submit your email on the home page . PS: SOAP use Google group as it's mailing list .

05-02-2011, 08:50 AM	#6
rubi Junior Member Location: North Carolina Join Date: Apr 2011 Posts: 6	Hi, Can anyone refer me to good documentation on quality assessment of soap denovo assembly (e.g., n50 values, how to compute coverage, and what the output files mean, etc...) Thanks

10-03-2011, 03:00 PM	#7
y.divyatej@gmail.com Junior Member Location: phoenix Join Date: Feb 2011 Posts: 4	Denovo assembly pipeline I'm curious if there is a pipeline available for Soap De novo assembly. Is there a requirement for the number of genomes required for Denovo assembly?

10-13-2011, 02:14 AM	#8
narain Member Location: Washington DC Join Date: Aug 2011 Posts: 78	Dear Members I am doing de-novo assembly of human genome from fastq data files. I get contigs as well as scaffolds from tools that I use. I know that scaffolds are a combination of contigs with estimated gaps in between them. Does this mean that downstream analysis when comparing it to another genome such as the reference should be done with contigs more reliably than with scaffolds ? Aby

06-13-2012, 08:22 AM	#10
sagarutturkar Member Location: Tennessee, USA Join Date: Sep 2010 Posts: 61	I am confused about contig numbers reported by Soap.I was trying to run SoapDenovo for small microbial genomes (Size varies from 5-10MB). However, the number of contigs reported in .contig files is very high (always in thousands) whereas other assemblers giving me contigs less than 500. But Soap-Scaffolding output was better than other assemblers. I am not sure, if I am looking at some intermediate contig file OR contig number is always high in Soap? From my experience contig and Scaffolds numbers differ by few 100s only (at least for small microbial genomes). There is no drastic change, but in Soap contigs were in 2000-3000 range while scaffolds were in 200-500 range. Please explain.