short reads Assembly in contigs

[mido1951]

Hello,
I made assembling illumina short reads of size 2 * 300bp and I used Velvet.
after the assembly, I got contigs with size less than the initial reads?
is it normal?
Thanks for your help

[GenoMax]

Is this the E coli dataset from Illumina you were asking about a few days back? Did you do any QC/cleanup of the data before assembly? Did you use default parameters for velvet?

[mido1951]

No, not E.Coli I have not yet done tests on E.Coli but I think I'll get the contigs less long than short reads.it's another organism.
Yes I did QC / cleanup and I assembled the raw data it's the same results. and yes I used the setting of velvet defects.
do you think that's normal?

velveth Assem 31 -shortPaired -fastq -separate /illuminaQCtool/W303_Miseq_R1.fastq_filtered /illuminaQCtool/W303_Miseq_R2.fastq_filtered
time velvetg Assem

[GenoMax]

How many contigs did you get? Are most of them long?

[mido1951]

I 2918681 contigs and different lengths. The longest contigs measure 1328. but others are shorter ~ 100bp.

[GenoMax]

Is this dataset not very good? Are you using published data and trying to recreate the analysis?

[mido1951]

yes I use data published in their raw format.

[GenoMax]

When you "filtered" data did you make sure R1 and R2 reads were kept in the matched order. What trimming/filter program did you use?

My guess is your filtered files must have gone out of sync as far as R1/R2 read order is concerned.

[mido1951]

I used the QCtoolkit filter program. and yes I have ensured that R1 and R2 are kept in good order.
Even with the raw data I got the same thing.

[GenoMax]

Do you have the reference available for this paper?

[mido1951]

not, but these are data that were used to make the assembly.

[GenoMax]

If this data was only used for correction then it may not have deep enough coverage for an independent assembly. Check into how much data there is and compare it to expected genome size.

[mido1951]

this is the link of QCtoolkit.
QCtoolkit