PE & Merge

Question

[Arash]

Hi, Before mapping , we have to merge PE reads?If yes, what's the way? Or doing map separably?

[GenoMax]

You can only merge PE reads if they actually overlap. That is generally possible when the number of cycles of sequencing is more than (size of the insert divided by 2) (for common applications, there are exceptions). You have seen some examples of that being noted below. This may also happen inadvertently if size selection/library prep is not done properly.

Answer 1

[WouterDeCoster]

In most typical analysis pipelines for standard applications you shouldn't merge the reads and your aligner will make good use of having PE reads. Most, if not all, common aligners will take two files (or more) simultaneously so you also don't have to map these separately.

[fanli.gcb]

For amplicon sequencing, PE reads are generally merged prior to clustering or what not.

[Arash]

Hi,thanks...would you please introduce me an aligner in windows?

[GenoMax]

Use BBMap. It is pure Java and will run on Windows. That said you would still need enough RAM to make it run (e.g. for human data significant 10+G free RAM needed). So make sure you have adequate hardware available.

That said I will concur with others that you would be better off finding central compute resources (if available at your institutions) to do this type of analysis. Renting cloud compute resources may also be an option if your budget allows for that.

[zizigolu]

once I have tried BBMap, really a good tool.

CUTADAPT in windows

specially in the mid of post I asked Brian about alignment

[WouterDeCoster]

No I can't and won't. You'll make your rest of analysis very painful if you decide that you will stick to windows. There is no good reason not to go for Linux. Using virtualbox or dual boot solutions you can also have the advantage of both operating systems. But setting that up is beyond the scope of this forum and you can find great guidelines online.

[GenoMax]

Virtual OS solutions don't make sense if the base hardware OP has is inadequate. When someone asks specifically for a windows solution it may not be out of bounds to assume that they have hardware that may barely meet the needs to run native analysis let alone a virtual OS (e.g. 8 or a max of 16G may be typically available).

It is easy to overlook the fact that many folks from less developed parts of the world do not have access to central compute resources that we may take for granted at a university/institution big or small. Personal hardware they may be trying to do the analysis on may not be the latest either.

[WouterDeCoster]

You're definitely right about hardware requirements, but I didn't make assumptions about what OP has access to (debatable if I should). I'm no sys admin and indeed have the luxury of central compute resources in my institution. I also didn't make assumptions about what exactly OP is trying to map, whether it would be amplicon sequencing, resequencing a lambda phage genome or whole genome sequencing on an octaploid plant species.

As a side note, if I'm not mistaken it is possible to have a windows machine running a VM on a compute cluster.

[Farbod]

Dear Arash, @WouterDeCoster is totally true about using Linux.

If you live in Tehran there are some good classes there such as Iran Linux House that can make you expert in the command line skills.

[zizigolu]

in this post you will find aligning in windows although linux is not comparable. also once i imaged a bio-linux on dvd that has bowtie already installed and you can try biolinux.

C: please some one help me step by step i was really exhausted

good luck