Do I need to trim adapters and remove PE contaminators for mate-pair libraries before using ALLPATHS

Question

[zhangdezhi008]

Dear all,

I am trying to use ALLPATHS-LG to denovo assemble a bird species genome (~1.1Gb). I prepared 3 libraries, including 350 bp insertion PE libary (60Gb reads), 3kb MP library (30Gb reads) and 8kb MP library (30Gb reads), all are 150 bp PE sequenced.

As you can see, I did not prepared a fragment library for ALLPATHS, so I used soapdenovo2 first, but it always get a much bigger assembly, then I moved to MASURCA, but it always failed before the completion (CA failed). Therefore, I am going to add another 200 bp insertion library and try Allpaths. Hopefully it can work.

Seems like MASURCA need not to trim adapters and remove PE contaminators for mate-pair libraries, do I need to trim adapters and remove PE contaminators for mate-pair libraries before using ALLPATHS?

Best regards,

Dezhi

[Gabriel R.]

what do you mean by "pe contaminators"? chimeric sequences?

Answer 1

[h.mon]

Unless the assembler removes the adapters (and ALLPATHS-LG doesn't remove adapters), you always have to remove adapters for assembly.

As for the mate pairs, processing them (for example, with NxTrim) should result in true mate pairs, unknown pairs, and paired-end contamination. You can use the the unknown pairs as mate pairs, most of them are mate pairs but the biotinilated adapter was not found because the insert was larger the sequence length. The paired-end contamination can be used as a regular paired-end library.

I don't think you need an additional 200bp library, it will add little to the three libraries you have.

Seems like MASURCA need not to trim adapters and remove PE contaminators for mate-pair libraries

Why do you think so? Does the MaSuRCA documentation says so? I am asking because I am not familiar with MaSuRCA.

[zhangdezhi008]

Thanks for your reply. I made a mistake, it was not adptors for MP library, it should be juctions of MP reads. So juctions should also be removed before using ALLPATHS, right?

[h.mon]

Yes, the biotinilated adapter (which I guess is what you are calling "junctions of MP reads") should be removed, NxTrim does this.