Data Sets

Teaser can benchmark mappers for simulated and real data. The main difference is that for simulated data, Teaser uses the source position of each read as given by the simulator, as gold standard for evaluating if a read was mapped correctly. For simulated data, you can either choose to use the built-in simulation pipeline (Mason or DWGSIM) or to import a custom simulation.

Data sets can be created using the Web Browser Interface or by manually creating the Configuration Files.

Simulated Data

A: Simulated using Teaser

In this case, you set the simulation parameters, which Teaser will use for reference downsampling and read simulation. For a list of simulation parameters see Table of Simulation Parameters.

B: Custom Simulation

Teaser is in principle compatible with any simulator that outputs reads in .fastq format and gold standard alignments in .sam format. The .sam file should contain a row for each read, with query name, flags, chromosome and position set accordingly. Place custom simulation files in the import directory.

Real Data

When benchmarking real read data, Teaser first samples reads from the supplied .fastq file(s). It then evaluates mapped rate (percent) and performance for each mapper. Of course, correctness of alignments cannot be evaluated for real data. Place real read data files in the import directory.

Further Information

For further information on how we create and evaluates data sets, see the Teaser publication.

Configuring Data Sets

Simulated, custom simulated and real data sets can be created/imported using the web interface. Alternatively, you can define them in configuration files.