|
Karma Options Reference
Mapping Controls
| Karma Option |
Description |
| --pairedReads | By default, Karma will map each read file independently. If this option is specified, reads will be taken two at a time, from the corresponding line offset in pairs of files. Information about the relative location of match candidates for the pair will be used to map the reads and in many cases provide a better map. |
| --seed [int] | A random number generator is used to break ties when map candidates of equal quality are found. This option sets the seed for it. |
| --prefix [pathname] | Each output file gets prepended with the prefix set here. By default it is "./", the current working directory. It should probably be named "outputPrefix". |
Read Filtering Controls
| Karma Option |
Description |
| --misMatchCutoff [int] | Drop a read when more than this number of mismatches is found in the first 30 bases of the read. The indexing algorithm allows zero or one mismatch in one of the index words, and up to this matchmatch cutoff total in the other word. Currently, mismatches in the remainder string following the first 30 bases are not counted in this computation |
| --maxInsert [int] | For paired reads, the distance between the outermost bases of the read pair can be limited by this paramater. This paramater is very process dependent. The default is 1000 bases. |
| --minimumMapQuality [int] | Single or pairs of reads with this map quality or lower are dropped. Single and paired reads are assigned a map quality between 0 and 100 inclusive. A map quality of 100 means a perfect match in the first 30 bases. |
| --maxReads [int] | Limit the number of reads to this paramater. Used to sample for example the first 10,000 reads of a file. If the value is 0, all reads are processed. The default is 0. For paired reads, each end is counted as a read. |
Read Trimming Controls
| Karma Option |
Description |
| --trimLeft [int] | Ignore the first [int] bases of the read. This is sometimes useful if it is known that for example the first 5 bases in a read are bad. The default value is zero, which means to not ignore any bases in the read. |
| --trimRight [int] | Ignore the last [int] bases of the read. This is sometimes useful if it is known that for example the last 5 bases in a read are bad. The default value is zero, which means to not ignore any bases in the read. |
Chromosome Reference Controls
| Karma Option |
Description |
| --reference [pathname] | The genome reference is required to map reads. This option sets the original fasta name of the reference (not the binary versions). The default for this is ideally a file common to all users on the machine, since then the genome and more importantly the genome word index will be efficiently shared by all users. The file may contain one or all chromosomes of a genome. |
Chromosome and Index Creation Controls
| Karma Option |
Description |
| --createIndex | Create the both the binary version of the reference and the binary word index for the reference chromosomes contained in the file --reference. |
| --createReference | Create just the binary version of the the reference chromosomes contained in the file --reference. This is useful if you are using our libcsg library to process the genome. |
Map File Contents Output Controls
| Karma Option |
Description |
| --showReferenceBases | When writing mapped records in the SAM output file, show the mapped string as a full set of bases, with capital letter bases indicating a match, and a lower case base letter as a mismatch. This is intended for debugging or other work. It could better be replaced with a more general option indicating human readable output. |
Other Karma Options
| Karma Option |
Description |
| --debug | Enables some additional debug steps internally as well as debug output. |
| --maponeread [read string] [quality string] | An experimental and mostly untested option for mapping a single read string. It may be useful in some circumstances. |
| --test | Run additional test code. This runs, among other things, slow code to verify the integrity of the word index. The intention is to eventually run more in-depth unit tests in karma. |
|
