FAQ

Q1:Does ARDEN support paired end reads?

There is no native support for paired ends. It is possible to treat these as single end reads and adjust the analysis.

Q2: What is the recommended distance parameter for creating the artificial reference?

There is no optimal value. The distance (d) parameter strongly depends on your application (error level) and your read length. It should be chosen according to your error level (e.g. read length= 100, error level= 5%, a reasonable choice would be d=21.

Q3: I want to compare a lot of different mappers, how can I decrease the runtime?

The initial implementation has no support for parallel analysis. It is however possible to split the input file accordingly in x-parts. The result files need to be put together manually.

Q4: Mapper x cannot be sorted with SAM tools, how can I fix this?

Contact the authors. If there is something wrong with the output format of mapper x it may also result in wrong computations within ARDEN.

Q5: How can I create artificial references for several chromosomes?

At the moment there is no support for this kind of input. ARDEN expects the input fasta file to have no "N"s and have a single entry per fasta file.

Q6: How can I analyze multiple artificial references?

To include multiple artificial references (A) with e.g. varying distances additional lines in the control.ini file are necessary. For each additional A an extra line starting with '#' has to be added to the header. In the same ordering the alignments in sam format have to be added. Such .ini file could look like this:


$:/data/test/reference.fasta
#:/data/test/artificial_1.fasta
#:/data/test/artificial_2.fasta
&:/data/test/reads.fastq
@MyMapper
ref:/data/test/reference.sam
art:/data/test/artificial_1.sam
art:/data/test/artificial_2.sam
+


Q6: How do I compute the best cut-off?
To find a good cut-off it is required to have a look at the generated ROC tables. Since ARDEN provides detailed information about sensitivity, specificity and AUC it is up to the user to choose an appropriate error level. Sorting the ROC tables, by PTP and FP might also give a good overview for suitable cut-offs.

If your question was not covered here do not hesitate to send an email.