Angsd structure

This page contains a short description of the overall structure of angsd.

It can be divided into:

1. Filereading
2. Analysis

We allow for many different input formats, many of these are deprecated and have little use, but we have decided to keep these.

1. BAM files
2. soap alignment files
3. Beagle files
4. glfv3 (Binary and text)

The data for the different input formats are encapsulated in a struct 'funkyPars', and this struct is sent to the 'analysis' classes, which will do analysis and populate the struct with results (eg based on the sequencing data we calculate genotype likelihoods and attach this to the struct).

Depending on the input format the information in the struct will span different regions. For the Beagle files the region size is defined by -chunkSize. The region size for BAM input can very quite alot since this is dependent on number of reads in the different BAM files. A thorough description of the BAM reading can be found in BAM reading details

Analysis

We have defined the following abstract base class called general.h

class general{
public:
  static aHead *header;//contains the header of a single bam;
  static std::map<char *,int,ltstr> *revMap;
  int index;
  static int tot_index;
  //  virtuel general()
  virtual void run(funkyPars *f)=0;
  virtual void print( funkyPars *f)=0;
  //  virtual void printArg(const char *fname)=0; <-maybe include
  virtual void clean(funkyPars *f)=0;
  general(){index=tot_index++;};
  virtual ~general(){};
};

History

Angsd started back in 2009 as a simple program for estimating allele frequency, based on this we incrementally added new functionality. A rough timeline can be found in the AUTHORS file in the program bundle