Assembly for a single cDNA

Question

[karpet34]

Dear all,

I have performed a RT-PCR which give me a 1500 pb product. I sequenced it with the Illumina technology (2X250 paired-end reads). Then, since several weeks, I unsuccessfully assemble the reads to get the full length sequence. I have tried many of classic assemblers (cap3, ssake, arapan, minimus2, ...) but all of them provide multiple contigs some of which exceeds more than 5kb!

I checked that all the reads are mapped well on the reference.

I am looking for an assembler able to do the job. Is there anyone have an idea?

Thank you!

[karpet34]

Hello, genomax was right: I got too much reads. Normalization with bbnorm and assembly with cap3 provide me a very good result. Now I must find the good option values to get a perfect assembly. Thank you all!

[lieven.sterck]

A small educational note: if an answer was helpful you should upvote it, if the answer resolved your question you should mark it as accepted.

[Ram]

What is the organism you're working on? You mention both align to reference and assemble. Do you wish to do both, and if so, why?

[karpet34]

Thank you for your answers. I will rework according to your advices. I remove bad quality reads (nearly no reads removed), remove Illumina adapters, merge read 1 and 2 then I proceed to assembly. I align to a reference because only a part of my cDNA is known. The 3' part is unknown so I need to perform de novo assembly of this part.

[lieven.sterck]

is the reference you're talking about genomic or also transcriptomic?

Answer 1

[GenoMax]

You probably have way more coverage than you need and that is likely causing the assembly problems. So consider downsampling the data. You can use bbnorm.sh from BBMap suite (guide here). You can also take a look at tadpole.sh (guide here) as an alternate k-mer based assembler.

Answer 2

[h.mon]

Illumina sequencing is noisy, because due to the sheer volume of data generated, even low frequency errors and contaminants will get a good number of reads in the end. When assembling an amplicon, you will want to filter your contigs by coverage, as your "true" amplicon will have a much higher coverage than the errors and contaminants. Also, there are several pipelines specialized for targeted sequencing assembly, you may try one (or several) of them. Two I remember are ARC and HybPiper.

By the way, did you remove Illumina adapters before assembling?