Genome assembly and Rna-seq mapping

[freestile]

I every one, is my first time here but I follow for many months. My question is if someone have experience with genome assembly and post mapping of ran-seq data.
I use velvet, celera and clc to make de novo assembly of relative difficult bacteria, and i have relative goods statistics with velvet and celera (revised with published draft genomes). Then I check the assemblies mapping rna-seq data of the same organism and I obtain lows mapping percentages (near 50%) but when I check clc assembly this up to 90%. I try to optimise the clc assembly but if I drecrease the data or make more stringent trimming, the assembly decrease in quality.

I use 250pb PE in miseq plataform.

Best regards !
Cristian.

[SylvainL]

Hi,

for bacteria, I was using Edena http://http://www.genomic.ch/edena.php with good results

Maybe you could give a try...

[freestile]

Hi,

for bacteria, I was using Edena http://http://www.genomic.ch/edena.php with good results

Maybe you could give a try...

Thxs for reply. I will try. I obtain high rna-seq reads mapping with a5 pipeline, but i can't increase genome assembly statistics.

[freestile]

Hi,

for bacteria, I was using Edena http://http://www.genomic.ch/edena.php with good results

Maybe you could give a try...

I try Edena but i can't get the same length of my reads, If you can help me I appreciated.

[SylvainL]

Hi,

what do you mean by not getting the same length of your reads? In my case, when the Qscores were bad at the end of the reads, I had to trimm them... You can easily do this step with fastx toolkit (fastx_trimmer in this case) or seqtk (trimfq)...

Let me know at which step you are stuck...

[freestile]

Hi,

what do you mean by not getting the same length of your reads? In my case, when the Qscores were bad at the end of the reads, I had to trimm them... You can easily do this step with fastx toolkit (fastx_trimmer in this case) or seqtk (trimfq)...

Let me know at which step you are stuck...

The error is when start edena:

Rapid file(s) examination... 158 220
[err] All reads within a file must be the same length.

I make pre-processing with bbmap. Maybe the problem is paired end data?

[fahmida]

I am also having the same error "[err] All reads within a file must be the same length". I am not sure I understand the rationale behind this. Once the raw reads e.g. 100 or 150nt length are trimmed for quality, adapters etc., read length becomes variable.

[GenoMax]

For bacterial assemblies SPAdes (http://bioinf.spbau.ru/spades) should be in your list of programs to try.

[freestile]

I try spades, and really can't obtain good assembly in my specific data (I try with other bacteria data and I had got good results). This is the principal reason why I try others assemblers.

[SylvainL]

I am also having the same error "[err] All reads within a file must be the same length". I am not sure I understand the rationale behind this. Once the raw reads e.g. 100 or 150nt length are trimmed for quality, adapters etc., read length becomes variable.

In this case, I would advice you to look at the minimum length of your reads and trim all of them to have this minimum length, or depending of your FastQC report, I would go without adapter trimming... Really it is worthwhile trying Edena. As example for a total de novo assembly of Staphylococcus aureus, I got 12 contigs (which stopped because of the rRNA operons).

s.

[freestile]

Thxs for your help. Finally I run edena, but I can't improve rna-seq mapping (~50%) comparing with clc, a5 and spades (~90%). I think that the next step is choice assembly with relative good statistics and at least ~90% rna-seq map.

Regards.

[SylvainL]

Thxs for your help. Finally I run edena, but I can't improve rna-seq mapping (~50%) comparing with clc, a5 and spades (~90%). I think that the next step is choice assembly with relative good statistics and at least ~90% rna-seq map.

Regards.

When you say your rna-seq mapping is low do you mean only 50% of your rna-seq reads are mapped to your assembly? If yes, did you try a blast on some unmapped reads to see if they really come from your bacterium (or closed to)?

Another question: after rna-seq mapping to your asembly, do you have some regions which are not covered at all?

These RNAseq were performed with rRNA depletion? Do you have rRNA operons on your assembly?

[freestile]

When you say your rna-seq mapping is low do you mean only 50% of your rna-seq reads are mapped to your assembly? If yes, did you try a blast on some unmapped reads to see if they really come from your bacterium (or closed to)?

Another question: after rna-seq mapping to your asembly, do you have some regions which are not covered at all?

These RNAseq were performed with rRNA depletion? Do you have rRNA operons on your assembly?

Yes, only 50% and I blast some unmapped reads and correspond to bacteria genes.

I have to check te second question

And yes I make rRNA depletion in library preparation.

Regards !