One way of measuring contamination or filtering out reads that most likely originate from modern DNA is to use the post-mortem damage score (PMDS) introduced by Skoglund et al. (2014). With this task, you can use ATLAS to calculate the PMDS for all reads. The output is a new BAM file, where each alignment is described by a field DS:f:<PMDS>, which specifies the alignment's PMDS score. If the DS field already exists it will be overwritten. It is possible to filter out the reads with a low PMDS score while writing the BAM file by using the option minPMDS or maxPMDS to specify the threshold value.
A PMDS score > 0 means that the read is more likely ancient than modern. The higher the score, the higher its probability of being ancient.
- A BAM file (filtered or not), where each alignment has a tag DS:f:<PMDS> recal=example_recal.txt
./atlas task=PMDS bam=example.bam pmdFile=example_PMD_input_Empiric.txt recal=example_recalibrationEM.txt verbose
- minPMDS : filter out reads that have a PMDS score below the specified value. Default = -10000
- maxPMDS : filter out reads that have a PMDS score above the specified value. Default = 10000
Engine parameters that are common to all tasks can be found here.