Here's a Python script which converts BAM files to wiggle, then BigWig format.

This uses pysam to get a pileup via samtools -- essentially the same approach as your one liner -- so the speed will be comparable. The script allows specification of specific chromosome regions so is useful for examining a region in depth without having to skip over any bases.

We upload BigWig files to a local Galaxy instance. It can serve out sections of the file on-demand, which gives you the advantage of wiggle display at UCSC without the upload times.

For me wiggle representation is generally for the coverage, so I'll just compute the coverage first using bedtools utility as genomeCoverageBed -i file.bed -bg -g my.genome > sample.cov, where file.bed can be obtained using bamToBed utility bamToBed -i file.bam > file.bed and then would use bedGraphToBigWig utility for ucsc to convert the coverage file to bigwig, which is better than wig as you can place on the local server and pass the text file containing links of these bigwigs to collaborators etc. for viewing with UCSC and its safe as only you having rights can change the file content. Another option would be using pyicos convert tool as pyicos convert my_experiment.bam my_experiment.wig -f SAM -F variable_wig -O, this will give you wig directly.

Cheers

I've been using Findpeaks which is actually a ChIP software, but there's an option (-dist_type 3) to make it only count reads...which is what you would want.

It can also break down the track on a per chromosome basis, or one track with all the chromosomes.

http://sourceforge.net/apps/mediawiki/vancouvershortr/index.php?title=FP4Parameters

veggie's bam2ucsc might be useful to visualize it in the ucsc browser.

Script can be found here.

In recent versions, the samtools pileup command has been removed

[main] The 'pileup' command has been removed. Please use 'mpileup' instead

So, I found a solution that uses samtools depth instead

samtools depth fileName.bam | perl -ne 'BEGIN{ print "track type= print wiggle_0 \
name=fileName description=fileName\n"}; ($c, $start, $depth) = split; \
if ($c ne $lastC) { print "variableStep chrom=$c span=10\n"; };$lastC=$c; \
next unless $. % 10 ==0;print "$start\t$depth\n" unless $depth<3' > fileName.wig

It isn't necessarily faster or more elegant but it works with latest versions of samtools. Also, there's a similar thread here that talks about using samtools mpileup: convert bam to wig

Edit: If you are converting to bigwig and receive the error "chromosome xxx has nnnn bases but ends at nnnn", then it's probably because the added span parameter goes beyond the end of the genome. Therefore, use the -clip option when doing wigToBigWig

Edit 2016: I would use bedtools coverageBed and then bedGraphToBigWig instead of this method

Is there anyway to make pileup include duplicates marked reads in the coverage?

The python script showed above doesn't work well for me. So I would recommend samtools mpileup as well.