Input format

BAM file

Mclumi receives a bam file as input by default after using a trimmed *.fastq.gz file with each read umi attached to the read name. In single cell analysis, barcodes will be attached to the read name prior to their UMIs. The *.fastq.gz can be mapped on a reference genome as a bam file using a mapping tool, such as HISAT2 or STAR. For simulation purposes, the bam file can be generated using another pipeline called simReadFlow, which uses a fastq.gz file as input directly and outputs a bam file because simReadFlow will control the format of simulated fastq.gz files to be recognized by trimming module.

UMI-tools preprocessed BAM file

Besides, we noticed that a well-curated module for preprocessing a BAM file is embedded in UMI-tools. To enable use of this UMI-tools module in Mclumi, you can import a module BundlePos with its class object bundlePos to deal with this issue. All qualified reads will be accessible to section tags PO='pos' in the bundle bam file. Unqualified reads are filtered by UMI-tools modules, due to factors such as poorly recognized contigs.

from mclumi.align.BundlePos import bundlePos
in_fpn = to('your_working_path/example.bam')
out_fpn = to('your_working_path/example_bundle.bam')
bundlePos.convert(options=options, in_fpn=in_fpn, out_fpn=out_fpn)