A Python API for RNA-seq alignment, quantification normalization via Kallisto. PyKallisto can take a list of raw fastq files in a compressed or uncompressed format, and generate a cell x gene expression matrix and/or
a cell x transcript equivalent class count matrix, quantify data from many common technologies such as 10x, indrops and dropseq, process feature barcoding data, and obtain QC reports for single-cell rna-seq data. Under the hood, PyKallisto is entirely reliant on kallisto, bustools, and kallisto | bustools from the Patcher Lab.
All thanks for the technologies and beautiful command line tools goes to them. All I did was wrap them in a Python API for easy handling of data in larger pipelines, to allow for pure Python single-cell analysis in a single line of code.
pykallisto has two classes, Kallisto and KallistoBus. The former wraps kallisto, and the latter wraps the kb | tools from the Patcher Lab.
For both tools, the arguments list are the same as the extended argument list in the command line version. That is, instead of -o/--output-dir for the output directory, we pass output_dir=/path/to/dir in the Python classes.
Example:
from pykallisto import Kallisto
data = Kallisto(files=['fastq1.fasta', 'fastq2.fasta', 'fastq3.fasta'])
# Generate the index file for pseudoalignment
data.index(index='index', output_dir='index_directory/')
# Quantify transcripts in parallel on 4 threads
data.quant(output_dir='results/', threads=4)