Trim

The Mclumi Trim module allows users to trim UMI(s) and/or barcode(s) of any length as well as other components (e.g., primer) from a complex read structure, which is much more flexible than other UMI tools. For instance, it can be used for trimming from reads sequenced by template bulk-RNA-seq and scRNA-seq and so on. The module works based on a clear defined structure of a read. Therefore, the module needs the read length to be indentical for all as well as UMI-tools does. In python inline mode, a set of parameters need to be specified in json format for the composition/structure of your input reads. For example, an input read consists of a barcode first, a umi then, and a read finally. Details of an exmaple of reads by template switching oligos (TSO) are given below.

Python inline

{
    # umis
    'umi_1': {
        'len': 12,
    },
    'umi_2': {
        'len': 12,
    },
    # ...,

    # reads
    'seq_1': {
        'len': 30,
    },
    'seq_2': {
        'len': 30,
    },
    # ...,

    # other components, e.g.,
    'bc_1': {
        'len': 20,
    },
    # ...,

    'read_struct': 'bc_1+umi_1+seq_1+umi_2',
    'fastq': {
        'fpn': 'xxx.fastq.gz',
        'trimmed_fpn': 'xxx_trim.fastq.gz',
    },
}

from mclumi.trim.Fixed import fixed
p = fixed(
    mode='internal',
    params=params,
)
p.call()

As an example, we take a simulated fastq file of small size, which can be download via pcr_1.fastq.gz.

If you tend to run Mclumi in inline mode, the parameter of mode should be specified as internal. This will tell Mclumi to run the Trim module inline. After running starts, the module pops out prompts like this

22/10/2021 20:03:28 logger: run Mclumi internally.
22/10/2021 20:03:28 logger: Your params for trimming UMIs are:
{'umi_1': {'len': 12}, 'umi_2': {'len': 10}, 'umi_3': {'len': 12}, 'primer_1': {'len': 20}, 'primer_2': {'len': 20}, 'seq_1': {'len': 6}, 'seq_2': {'len': 8}, 'read_struct': 'primer_1+umi_1+seq_1+umi_2+primer_2', 'fastq': {'fpn': 'pcr_1.fastq.gz', 'trimmed_fpn': 'pcr_1_trim.fastq.gz'}}
22/10/2021 20:03:28 logger: ===>reading from fastq...
22/10/2021 20:03:28 logger: ===>umi structure: primer_1+umi_1+seq_1+umi_2+primer_2
22/10/2021 20:03:28 logger: ===>umi positions in the read structure: 1, 3
22/10/2021 20:03:28 logger: ===>seq positions in the read structure: 2
22/10/2021 20:03:28 logger: ======>finding the starting positions of all UMIs...
22/10/2021 20:03:28 logger: =========>umi_1 starting position: 20
22/10/2021 20:03:28 logger: =========>umi_2 starting position: 38
22/10/2021 20:03:28 logger: ======>finding the starting positions of all genomic sequence...
22/10/2021 20:03:28 logger: =========>seq_1 starting position: 32
22/10/2021 20:03:28 logger: ===>umi_1 has been taken out
22/10/2021 20:03:28 logger: ===>umi_2 has been taken out
22/10/2021 20:03:28 logger: ===>seq_1 has been taken out
22/10/2021 20:03:28 logger: ===>start saving in gz format...
22/10/2021 20:03:28 logger: ===>trimmed UMIs have been saved in gz format.

CLI

If you tend to run Mclumi in CLI, only quite a few differences should be made. -l the lengths of different components should also be concatenated by +.

mclumi trim -i ./pcr_1.fastq.gz -o ./pcr_trimmed.fastq.gz -rs primer_1+umi_1+seq_1+umi_2+primer_2 -l 20+10+40+10+20