Skip to content

zerodel/working_pipe_line

Repository files navigation

design convention

top-level interface

the name convention of top level pipeline/workflow should begin with "wf_*"

what should top level interface do ?

it should

1. load the config files
2. read the additional arguments
3. sub-level of these module should be tool sets divide by usage. such as align tools, quantification tools, DE-detectors. etc. the final tool should be decide by the sub-level tool sets.
4. top-level module should provide output check and input check for the sub-level module, that is pre-module function and post-module function.

input arguments or input file

bunch mode

hmm.. , a often seen job is so-called "bunch", that is , a bunch of reads file with the same index .

so we need a input file , with contains the locations of these reads, and a setting file, which stands for the setting for all reads in this bunch .

option checker

the config file remains mostly all the original cli options. a controller/checker should be established, and should provide following feature:

1.check the options before invoking external tools, show description message when error occurs
2.give some query/info interface

every module which contains a option_checker should have this constant OPTION_CHECKERS , which is a list of opt-checker objects

output / logging part

yes , the output file is the logging file, and should use the standard logging module. ... but still undone yet . it is mostly the python logging manual's fault .....

option checker and helper

a mechanism for checking opts and print help is needed .

QC tools

here, we assume the QC tool is fastqc.

and those qc tool should have following parts:

  1. run_qc

a function/method to invoke the QC pipeline

  1. parse_qc_result

a function to parse QC result

  1. SECTION_QC_SETTING

the name of QC setting section in configure file

  1. OPTION_CHECKERS

a list of option check , this will generate the default description string

aligner

most aligner need some kind of index to run the mapping.

so all aligner wrapper should have following feature:

  1. index

function interface for establishing index

input:
output: file_path of indice, which can be understand by the aligner itself

if the index already exists, just return the path.

  1. align

function interface for mapping sequence reads to reference genome.

input: sequence reads in fasta/fastq
output: return a dict contains serval entries ,
  1. is_path_contain_index

function for checking whether the given path have a index this is done by checking existence of related files. those related files should be provide by aligner module itself.

input : file path to possible index
output : bool value indiacting whether there is a index .
  1. SECTION_INDEX:

a string shows index parameter is under configure[SECTION_INDEX]

  1. SECTION_ALIGN:

a string shows align parameters are under configure[SECTION_ALIGN]

  1. interpret_index_path:

function : translate a given index path to a dict which is compatible with configure file section.

input: a string containing possible index path

output: a dict , which can be accept by ChainMap to make a dict as input of index/align

when input is None, return {}
  1. interpret_seq_files

function: translate string contains a path or paths of sequencing read file to a dict which is compatible with configure file section

input: a string contains one or more sequencing reads file path
output: a dict , which can be accept by ChainMap to make a dict as input of align function

if input is None, return {}
  1. is_map_result_already_exists

function for checking whether the alignment result already exists

input : file path, or path/prefix specify the path
output: bool value shows whether there already have mapped result . 
  1. get_align_result_path

function to get the path of mapped result file

input : the same input as the align function 
output: path or path/prefix showing where the output file should be 

suggested paradigm

def index/quant(config, ...)
    get_dict(...)
    check_opts
    get_cmd
    run_cmd

quantification tools

similar with aligners, quantification tools also need index . so the main framework of quantification tools is the same as the aligners , only the spell differs .

  1. index

function interface for establishing index

input: ... some massive dict
output: return the file_path of indice
  1. quantify

function interface for calculating the expression

input: sequence reads in fasta/fastq
output: return a dict including the path of expression file
  1. is_path_contain_index :

function for checking whether the given path have index

input : file path to possible index
output : bool value indiacting whether there is a index .
  1. SECTION_INDEX:

a string shows index parameter is under configure[SECTION_INDEX]

  1. SECTION_QUANTIFY:

a string shows align parameters are under configure[SECTION_QUSANTIFY]

  1. interpret_index_path:

function : translate a given index path to a dict which is compatible with configure file section.

input: a string containing possible index path

output: a dict , which can be accept by ChainMap to make a dict as input of index/quant
return {} if None is given
  1. interpret_seq_files

function: translate string contains a path or paths of sequencing read file to a dict which is compatible with configure file section

input: a string contains one or more sequencing reads file path
output: a dict , which can be accept by ChainMap to make a dict as input of quant function
return {} if None is given

Circular RNA detecting tools

may need some sequence reads as input , this will be aligner's output . seems this tool do not need a index phrase . or it can be done manually.

  1. detect

main function interface ,

input: sequence_file, path to sequence reads fils in fastq/fa format
output: a dict(or other container data type), this should contain a path will the report of circular RNA detection
  1. export_as_bed

function to transform detection report to .bed file type

input: a dict contains settings for this dectection job. 
output: 1. a bed file contains the information of circRNA detection.
        2. optional gene-mapping file
  1. SECTION_DETECT:

a string shows where is the setting of this detection tool in config file.

  1. interpret_seq_files

function designed for work with files assigned by arguments,

  1. which_external_aligner

function to tell whether a external mapping should be performed before the detecting phase . return "" if no aligner is needed . else return the config section name of aligner such as "BWA", "STAR", etc

  1. bed_out

[optional] file path to bed type result, summarized by module's to_bed function

bed file format convention

bed file of circRNA are comparable with the default bed file format described in http://www.ensembl.org/info/website/upload/bed.html

meaning of each column are : first 3 column are required field.

  1. chrom : name of chromosome

  2. chromStart : start position of the circRNA

  3. chromEnd : end position of the circRNA

  4. name : should follow pattern : rna_name@gene_name_or_id, string after the first @ symbol will be taken as description,

  5. score : number of junction reads detected by detection tools

DE tool

TODO: need give lazy option for aligners

TODO: detector module should have a flag show whether it need an external aligner

About

a crude pipeline for RNA-seq data analysis

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published