Yesterday I attended the annual "Potatoes in Practice" meeting for the first time, mainly to see the finished display which I helped produce. Here it is, showing the twelve chromosomes of potato, drawn as stylized uniform green 'X' shapes, with different colour LEDs marking traits of interest for potato breeding.

Mark-up languages allow you to write a plain text input file which is then processed to produce a nicely formatted output - as HTML, PDF or similar. The plain text nature of the input is perfect for tracking under version control software, which other richer formats are not suited to. So what's the current best markup choice for a Python project?

I've written about random access to the blocked GZIP variant BGZF used in BAM, and looked at random access to BZIP2, but here I'm looking at XZ files which are based on LZMA compression. This was prompted by the release of Python 3.3 which includes the lzma module to support XZ files, which I then back-ported to offer lzma for Python 2.6 or later.

Scientific software released for free under an "academic only"/"non-commercial" licence really irks me, even if the source code is included. This is mainly because it isn't open source, and I believe science should be open, but this also causes a whole lot of headaches.

The blastdbcmd tool in the BLAST+ suite (replacing fastacmd in the C 'legacy' BLAST suite) lets you do a lot of clever things with a BLAST database. As long as you follow the baroque NCBI FASTA naming scheme you can do this with local BLAST databases too. However, if you don't want to bow down to the NCBI naming (e.g. use FASTA files directly from your favourite assembler), then blastdbcmd seems needlessly crippled.

This is just a quick post to document a bug in the blastdbcmd tool from the BLAST+ suite when used on the NR database with a full identifier and the -target_only option.

One of the first things a programmer dealing with 'Next Generation Sequencing' (NGS) aka 'High Throughput Sequencing' (HTSeq) data learns is to be very aware of memory limitations. You can't just go loading files into RAM when they are often gigabytes in size. Instead where possible you loop over a file (iterating over it record by record) or employ indexed random access. The authors of MrFast & MrsFast didn't do this.

Today's release of Python 3.3 includes the lzma module in the standard library for  LZMA/XZ files, but it didn't work 'out of the box' under Mac OS X 10.8 Mountain Lion - it requires XZ Utils. This is how I installed it. At the time of writing Python 3.3 doesn't come pre-compiled for Mac OS X 10.8 Mountain Lion (only older versions of OS X), so I built it from source (having installed XCode via the App Store for free).

Have you ever tried to use a BLAST database of protein sequences containing stop codons? If you work on nice model organisms with solid gene annotation maybe not. However, with draft annotations, mutation studies, or read through translation it is not unreasonable for the odd internal stop codon to appear in a protein sequence. And some translation pipelines do leave in a trailing * character.

This is an open letter to the NCBI BLAST+ team to request two simple enhancements which I think would be extremely useful - first and foremost the option to include BLAST result descriptions in the tabular output. Having the taxonomic identifiers (if available) would be great too - allowing downstream filtering of BLAST results by species etc.

Sometimes using BLAST is frustrating. Today I'm writing about it returning different expectation values, and therefore different answers, depending on if you use a FASTA subject file, or a database made from that file.