Using Kmers And Read MAppings to optimize (de novo) transcriptome assemblies.
conda install -c bioconda -c conda-forge -c lmfaber karma
usage: karma.py [-h] -i STR [-1 STR] [-2 STR] [-s STR] [-o STR] [-t INT]
[-l {DEBUG,INFO,WARNING,ERROR,CRITICAL}] [--lowest]
[--rearrange] [--threshold THRESHOLD] [-d] [-k STR]
[--n_neighbors N_NEIGHBORS] [--n_components N_COMPONENTS]
[--min_dist MIN_DIST] [--random_state RANDOM_STATE]
[--min_cluster_size MIN_CLUSTER_SIZE] [--inflation INT]
[--annotate] [--busco-group STR] [--database-dir STR]
karma. Using Kmers And Read MAppings to optimize (de novo) transcriptome
assemblies.
optional arguments:
-h, --help show this help message and exit
KARMA:
Necessary arguments for karma.
-i STR, --input STR Input fasta file to cluster. (default:
None)
-1 STR, --left STR Left reads. (default: None)
-2 STR, --right STR Right reads. (default: None)
-s STR, --single STR Single end reads. (default: None)
-o STR, --output STR Output directory. (default:
karma_output)
-t INT, --threads INT Threads to use. (default: 6)
-l {DEBUG,INFO,WARNING,ERROR,CRITICAL}, --log {DEBUG,INFO,WARNING,ERROR,CRITICAL}
Set the logging level. (default: INFO)
--lowest Representative sequence selection. Per
default karma will select the sequence
with the highest shared read
information in one cluster. With this
option the sequence with the lowest
shared read information will be
selected as well. Increasing
redundancy but may increase overall
information. (default: False)
--rearrange After kmer and read graph based
clustering the clusters can be
rearranged based on shared reads.
(default: False)
--threshold THRESHOLD Threshold for rearranging groups.
(default: 0)
-d, --draw Draw the graphs before and after read
based clustering. May take a lot of
time. (default: False)
KMER:
Kmer based arguments required for UMAP (Dimension reduction) and HDBSCAN
(clustering algorithm). See https://umap-
learn.readthedocs.io/en/latest/parameters.html#
https://hdbscan.readthedocs.io/en/latest/parameter_selection.html
-k STR, --kmer STR Kmer size to perform kmer based
clustering. (default: 5p6)
--n_neighbors N_NEIGHBORS UMAP: This parameter controls how UMAP
balances local versus global structure
in the data. (default: 2)
--n_components N_COMPONENTS UMAP: allows the user to determine the
dimensionality of the reduced
dimension space we will be embedding
the data into (default: 10)
--min_dist MIN_DIST UMAP: Minimum distance between two
data points. (default: 0)
--random_state RANDOM_STATE UMAP: Initialization value for
reproducible results. (default: 42)
--min_cluster_size MIN_CLUSTER_SIZE HDBSCAN: Minimum cluster size.
(default: 2)
MCL:
MCL related arguments. See https://micans.org/mcl/
--inflation INT Sets the main inflation value to INT.
This value is the main handle for
affecting cluster granularity. Between
1.2 and 6. (default: 2)
Dammit:
Parameters for de novo gene annotation. See https://github.com/dib-
lab/dammit
--annotate Switch for annotating. (default:
False)
--busco-group STR Which BUSCO group to use. Should be
chosen based on the organism being
annotated. (default: metazoa)
--database-dir STR Directory to store databases. Existing
databases will not be overwritten.
(default: None)
python karma.py -i file.fasta -1 left.fastq -2 right.fastq -o output_dir -t 30 --annotate --busco-group nematoda --database-dir database_dir
If you do not wish to annotate leave out --annotate, --busco-group and --database_dir.