Config File
The config file is a YAML file that contains the configuration. Different runs can be configured. We recommend using one config file per MPRA experiment or MPRA project. However, in theory, many different experiments can be configured in a single file. It is divided into version (version of MPRAsnakeflow used), assignments (assignment workflow), and experiments (count workflow). This is a full example file with default configurations: config/example_config.yaml.
1---
2version: "0.5"
3assignments:
4 exampleAssignment: # name of an example assignment (can be any string)
5 bc_length: 15
6 alignment_tool:
7 split_number: 1 # number of files fastq should be split for parallelization
8 tool: exact # bbmap, bwa or exact
9 configs:
10 sequence_length: 171 # sequence length of design excluding adapters.
11 alignment_start: 1 # start of the alignment in the reference/design_file
12 FW:
13 - resources/assoc_basic/data/SRR10800986_1.fastq.gz
14 BC:
15 - resources/assoc_basic/data/SRR10800986_2.fastq.gz
16 REV:
17 - resources/assoc_basic/data/SRR10800986_3.fastq.gz
18 design_file: resources/assoc_basic/design.fa
19 configs:
20 default: {} # name of an example filtering config
21experiments:
22 exampleCount:
23 bc_length: 15
24 umi_length: 10
25 data_folder: resources/count_basic/data
26 experiment_file: resources/count_basic/experiment.csv
27 demultiplex: false
28 assignments:
29 fromFile:
30 type: file
31 assignment_file: resources/count_basic/SRR10800986_barcodes_to_coords.tsv.gz
32 # label_file: resources/labels.tsv # optional
33 configs:
34 default: {}
Note that the config file is controlled by a JSON schema. This means that the config file is validated against the schema. If the config file is not valid, the program will exit with an error message. The schema is located in workflow/schemas/config.schema.yaml.
Version Settings
Set the version of the MPRAsnakeflow this configuration is used for. This is important for future updates. The version is used to check if the config file is compatible with the current version of the workflow. If the version is not the same, the workflow will exit with an error message.
version:
description: Version of MPRAsnakeflow
type: string
pattern: ^(\d+(\.\d+)?(\.\d+)?)|(0\.\d+(\.\d+)?)$
skip_version_check:
description: Skip version check
type: boolean
default: false
- version:
A string like “0.2.0” or “1.2”. When the major version is “0,” the minor version should match with MPRAsnakeflow, e.g., “0.2.0” is compatible with MPRAsnakeflow 0.2.0, 0.2.1, or 0.2.2. When the major version is greater than 0, the major version must match with MPRAsnakeflow. For example, a config of “1.2.1” is compatible with MPRAsnakeflow 1.7 or 1.0.
Assignment Workflow
The assignment workflow is configured in the assignments section. The following settings are possible:
assignments:
description: Assignments to run with configurations
type: object
patternProperties:
description: name of the assignment
^([^_\.]+)$:
type: object
properties:
alignment_tool:
type: object
properties:
split_number:
type: integer
default: 1
tool:
type: string
enum:
- exact
- bwa
- bbmap
- bwa-additional-filtering
default: bbmap
allOf:
- if:
properties:
tool:
const: bwa
then:
properties:
configs:
type: object
properties:
min_mapping_quality:
type: integer
minimum: 0
default: 1
sequence_length:
type: object
properties:
min:
type: integer
max:
type: integer
required:
- min
- max
alignment_start:
type: object
properties:
min:
type: integer
max:
type: integer
required:
- min
- max
required:
- sequence_length
- alignment_start
- min_mapping_quality
required:
- configs
- if:
properties:
tool:
const: bwa-additional-filtering
then:
properties:
configs:
type: object
properties:
sequence_length:
type: object
properties:
min:
type: integer
max:
type: integer
required:
- min
- max
alignment_start:
type: object
properties:
min:
type: integer
max:
type: integer
required:
- min
- max
min_mapping_quality:
type: integer
minimum: 0
default: 1
identity_threshold:
description: Identity threshold is used to choose which alignments are worth trying to rescue.
type: number
default: 0.98
mismatches_threshold:
description: Threshold of mismatches we investigate if we should try to rescue.
type: integer
default: 3
verbose:
description: print which alignments were rescured and which could not be rescued
type: boolean
default: false
required:
- sequence_length
- alignment_start
- min_mapping_quality
required:
- configs
- if:
properties:
tool:
const: bbmap
then:
properties:
configs:
type: object
properties:
min_mapping_quality:
type: integer
minimum: 0
default: 30
sequence_length:
type: integer
minimum: 1
alignment_start:
type: integer
minimum: 1
required:
- min_mapping_quality
- sequence_length
- alignment_start
required:
- configs
- if:
properties:
tool:
const: exact
then:
properties:
configs:
type: object
properties:
sequence_length:
type: integer
minimum: 1
alignment_start:
type: integer
minimum: 1
required:
- sequence_length
- alignment_start
required:
- configs
required:
- tool
bc_length:
type: integer
BC_rev_comp:
type: boolean
default: false
linker_length:
type: integer
linker:
type: string
pattern: ^[ATCGNatcgn]+$
FW:
type: array
items:
type: string
minItems: 1
uniqueItems: true
BC:
type: array
items:
type: string
minItems: 1
uniqueItems: true
REV:
type: array
items:
type: string
minItems: 1
uniqueItems: true
adapters:
type: object
properties:
3prime:
type: array
items:
type: string
pattern: ^[ATCG]+$
5prime:
type: array
items:
type: string
pattern: ^[ATCG]+$
NGmerge:
type: object
properties:
min_overlap:
type: integer
default: 20
frac_mismatches_allowed:
type: number
default: 0.1
min_dovetailed_overlap:
type: integer
default: 50
required:
- min_overlap
- frac_mismatches_allowed
- min_dovetailed_overlap
default: {}
design_file:
type: string
design_check:
type: object
properties:
fast:
type: boolean
default: true
sequence_collitions:
type: boolean
default: true
default: {}
required:
- fast
- sequence_collitions
strand_sensitive:
type: object
default: {}
properties:
enable:
type: boolean
default: false
forward_adapter:
type: string
pattern: ^[ATCGN]+$
default: AGGACCGGATCAACT
reverse_adapter:
type: string
pattern: ^[ATCGN]+$
default: TCGGTTCACGCAATG
configs:
type: object
patternProperties:
^([^_\.]+)$:
type: object
properties:
min_support:
type: integer
minimum: 1
default: 3
fraction:
type: number
exclusiveMinimum: 0.5
maximum: 1
default: 0.75
required:
- min_support
- fraction
minProperties: 1
oneOf:
- required:
- linker_length
- required:
- linker
- required:
- BC
required:
- FW
- bc_length
- design_file
- configs
- alignment_tool
- NGmerge
minProperties: 1
For each assignment you want to process, you must give it a name like example_assignment. The name is used to name the output files.
- alignment_tool:
Alignment tool configuration that is used to map the reads to the oligos.
- split_number:
To parallelize mapping for assignment, the reads are split into
split_numberfiles. For example, setting it to 300 means that the reads are split into 300 files, and each file is mapped in parallel. This is only useful when using a cluster. When running the workflow on a single machine, the default value should be used. The default is set to1. (For technical reasons, when multiple assignments are defined, all will be set to the maximum defined in the config.)- tool:
Alignment tool that is used. Currently,
bbmap,bwa,bwa-additional-filtering, andexactare supported. Default isbbmap.- configs:
Configurations of the alignment tool selected.
- sequence_length (exact, bbmap):
Defines the
sequence_length, which is the length of a sequence alignment to an oligo in the design file. Only one length design is supported.- alignment_start (exact, bbmap):
Defines the start of the alignment in an oligo. When using adapters, you must set the length of the adapter. Otherwise, 1 will be the choice for most cases.
- sequence_length (bwa, bwa-additional-filtering):
Defines the
minandmaxof asequence_lengthspecification.sequence_lengthis the length of a sequence alignment to an oligo in the design file. Because there can be insertions and deletions, we recommend varying it slightly around the exact length (e.g., ±5). This option enables designs with multiple sequence lengths.- alignment_start (bwa, bwa-additional-filtering):
Defines the
minandmaxof the start of the alignment in an oligo. When using adapters, you must set the length of the adapter. Otherwise, 1 will be the choice for most cases. We also recommend varying this value slightly because the start might not be exact after the adapter (e.g., ±1).- min_mapping_quality (bwa, bwa-additional-filtering, bbmap):
(Optional) Defines the minimum mapping quality (MAPQ) of the alignment to an oligo. MAPQs differ between bbmap and bwa. For bwa: When using oligos with only 1bp difference, it is recommended to set it to 1. BBMap is better here, and we can use, for example, 30 or 35. For regions with larger edit distances, 30 or 40 might be a good choice. Default is
30(use bbmap).- identity_threshold (bwa-additional-filtering):
(Optional) Identity threshold is used to choose which alignments are worth trying to rescue. Default is
0.98.- mismatches_threshold (bwa-additional-filtering):
(Optional) Threshold of mismatches we investigate if we should try to rescue. Default is
3.- verbose (bwa-additional-filtering):
(Optional) Print which alignments were rescued and which could not be rescued. Default is
false.
- bc_length:
Length of the barcode. Must match the length of
BC.- BC_rev_comp:
(Optional) If set to
true, the barcode is reverse complemented. Default isfalse.- linker_length:
(Optional) Length of the linker. Only needed if you don’t have a barcode read and the barcode is in the forward read with the structure: BC+Linker+Insert. The fixed length is used for the linker after a fixed length of BC. The recommended option is
linkerby defining the exact linker sequence and using cutadapt for trimming.- linker:
(Optional) Length of the linker. Only needed if you don’t have a barcode read and the barcode is in the forward read with the structure: BC+Linker+Insert. Uses cutadapt to trim the linker to get the barcode as well as the start of the insert.
- FW:
List of forward-read files in gzipped fastq format. The full or relative path to the files should be used. The same order in FW, BC, and REV is important.
- REV:
(Optional) List of reverse-read files in gzipped fastq format. Files have to overlap the FW read by at least 10 bp (see
NGmergeandmin_dovetailed_overlap). The full or relative path to the files should be used. The same order in FW, BC, and REV is important.- BC:
List of index-read files in gzipped fastq format. The full or relative path to the files should be used. The same order in FW, BC, and REV is important.
- adapters:
(Optional) List of adapter sequences to trim after merging the reads. Can be specified with:
- 3prime:
(Optional) List of adapter sequence at the 3’ end that should be removed.
- 5prime:
(Optional) List of adapter sequence at the 5’ end that should be removed.
- NGmerge:
(Optional) Options for NGmerge. NGmerge is used to merge FW and REV reads. The following options are possible (we recommend using the default values):
- min_overlap:
(Optional) Minimum overlap of the reads. Default is
20.- frac_mismatches_allowed:
(Optional) Fraction of mismatches allowed in the overlap. Default is
0.1.- min_dovetailed_overlap:
(Optional) Minimum dovetailed overlap. Default is
10.
- design_file:
Design file (full or relative path) in fasta format. The design file should contain the oligos in fasta format. The header should contain the oligo name and should be unique. The sequence should be the sequence of the oligo and must also be unique. When having multiple oligo names with the same sequence, please merge them into one fasta entry. The oligo name is later used to link the barcode to the oligo. The sequence is used to map the reads to the oligos. Adapters can be in the sequence, and therefore
alignment_starthas to be adjusted.- design_check:
(Optional) Options for checking your design fasta file. The design file cannot have
[or], duplicated headers, and for best performance, sequences should not be identical.- fast:
(Optional) Use a simple dictionary to find identical sequences. This is faster but uses only the whole (or center part depending on start/length) of the design file. Cannot find substrings as part of any sequence. Set to false for more correct, but slower, search. Default is
true.- sequence_collisions:
(Optional) Check if there are identical sequences in the design file. Default is
true.
- strand_sensitive:
(Optional) If enabled, the reads are mapped to the oligos in a strand-sensitive way by adding unique adapters to both ends of the oligo reference as well as the FASTQ files. By default, this option is not enabled.
- enable:
(Optional) If set to
true, the strand-sensitive mapping is enabled. Default isfalse.- forward_adapter:
(Optional) Adapter sequence added 5’ of the oligo. Default is
AGGACCGGATCAACT.- reverse_adapter:
(Optional) Adapter sequence added 3’ of the oligo. Default is
TCGGTTCACGCAATG.
- configs:
After mapping the reads to the design file and extracting the barcodes per oligo, the configuration (using different names) can be used to generate multiple filtering and configuration settings of the final mapping oligo to barcode. Use <your_config_name>: {} to use the default values for the keys. Each configuration is a dictionary with the following keys:
- min_support:
A minimum number of same BC that map to the same oligo. Larger value gives more evidence to be correct. But can remove lot’s of BCs (depedning on the complexity, sequencing depth and quality of sequencing). Recommended option is
3.- fraction:
Minimum fraction of same BC that map to the same oligo. E.g.
0.7means that at least 70% of the BC map to the same oligo. A larger value gives more evidence to be correct. But can remove lots of BCs (depending on the complexity, sequencing depth and quality of sequencing). Recommended option is0.7.
Experiment workflow (including counts)
The experiment workflow is configured in the experiments section. Each experiment run (contains one experiment file with all replicates of an experiment). The following settings are possible:
experiments:
description: MPRA experiments to run with configurations
type: object
patternProperties:
description: name of the experiment
^([^_\.]+)$:
type: object
properties:
bc_length:
type: integer
minimum: 1
umi_length:
type: integer
minimum: 1
adapter:
type: string
pattern: ^[ATCGNatcgn]+$
data_folder:
type: string
experiment_file:
type: string
demultiplex:
type: boolean
default: false
label_file:
type: string
assignments:
type: object
patternProperties:
^([^_\.]+)$:
type: object
properties:
type:
type: string
enum:
- file
- config
assignment_file:
type: string
assignment_name:
type: string
assignment_config:
type: string
sampling:
type: object
properties:
prop:
type: number
exclusiveMinimum: 0
maximum: 1
total:
type: integer
minimum: 1
required:
- type
allOf:
- if:
properties:
type:
const: config
required:
- type
then:
required:
- assignment_name
- assignment_config
- if:
properties:
type:
const: file
required:
- type
then:
required:
- assignment_file
configs:
type: object
patternProperties:
^([^_\.]+)$:
type: object
properties:
filter:
type: object
default: {}
properties:
bc_threshold:
type: integer
minimum: 1
default: 10
outlier_detection:
type: object
properties:
method:
type: string
enum:
- rna_counts_zscore
times_zscore:
type: number
exclusiveMinimum: 0
default: 3
required:
- times_zscore
default: {}
min_dna_counts:
type: integer
miminum: 0
default: 1
min_rna_counts:
type: integer
miminum: 0
default: 1
required:
- bc_threshold
- min_rna_counts
- min_dna_counts
sampling:
type: object
patternProperties:
^((DNA)|(RNA))$:
type: object
properties:
threshold:
type: integer
minimum: 1
prop:
type: number
exclusiveMinimum: 0
maximum: 1
total:
type: number
minimum: 1
seed:
type: integer
required:
- filter
variants:
type: object
properties:
map:
type: string
min_barcodes:
type: array
items:
type: integer
minimum: 1
required:
- map
- min_barcodes
# entries that have to be in the config file for successful validation
required:
- bc_length
- data_folder
- experiment_file
- demultiplex
- assignments
- configs
- bc_length:
Length of the barcode. This is used to extract the barcode from the index read. The barcode is extracted from the first
bc_lengthbases of the index read. When no reverse read is given andadapteris not set teh exact length is used to extract the DNA BC from the FW read.- umi_length:
(Optional) Length of the UMI. This is used to extract the UMI from the index read. The UMI is extracted from the last
umi_lengthbases of the index read. Please provide if you use UMIs.- adapter:
(Optional) Adapter sequence in the FW read when no reverse read is given. This is used to trim the sequence and retrieve the BC using cutadapt.
- data_folder:
Folder where the fastq files are located. Files are defined in the
experiment_file. The full or relative path to the folder should be used.- experiment_file:
Path to the experiment file. The full or relative path to the file should be used. The experiment file is a comma separated file and is decribed in the Experiment file section.
- demultiplex:
(Optional) If set to
truethe reads are demultiplexed. This means that the reads are split into different files for each barcode. This is usefull for further analysis. Default isfalse.- label_file:
(Optional) Path to the label file. The full or relative path to the file should be used. The label file is a tab separated file and contais the oligo name and the label of it. The oligo name should be the same as in the design file. The label is used to group the oligos in the final output, e.g. for plotting.
insert1_name label1 insert2_name label1 insert3_name label2
- assignments:
Per experiments multiple assignments can be defined (naming them differently). Everey assignment name contains the following configurations:
- type:
Can be
fileorconfig.filemeans that you use a mapping file which is tab separated and gzipped. It contains in the first column the barcode and in the second column the oligo name. This file can be generated by the Assignment workflow. When usingconfigthis means that you are referring to a assignment that is specified in this config file.- assignment_file:
When using
fileplease insert the path to the assignment file (tsv.gz). When usingconfigplease set the name of the config previously described the assignment that should be used.- assignment_name:
When using
configplease insert the name of the assignment specified in the config file.- assignment_config:
When using
configplease insert the name config of theassignment_nameyou want to use.- sampling:
(Optional) Options Randomly removing barcodes in the assignment. Just for debug reasons.
- prop:
Sample down the BCs in the assignment file to this proporion.
- total:
Sample down the BCs in the assignment file to this number.
- configs:
Each experiment run can have multiple configurations including filter and sampling options.
- filter:
(Optional) Filter options. These options are available
- bc_threshold:
Minimum number of different BCs required per oligo. A higher value normally increases the correlation betwene replicates but also reduces the number of final oligos. Default option is
10.- min_dna_counts:
Mimimum number of DNA counts per barcode. When set to
0a pseudo count is added. Default option is1.- min_rna_counts:
Mimimum number of RNA counts per barcode. When set to
0a pseudo count is added. Default option is1.- outlier_detection:
(Optional) Outlier detection. Methods and strategies to remove outlier barcodes in the final counts. The following options are possible:
- method:
Method to remove outliers. Currently
rna_counts_zscore,ratio_madornone(no outlier detection) are supported. Default option isrna_counts_zscore.- mad_bins:
(Optional) For method
ratio_mad: Number of bins for the median absolute deviation (MAD) method. Default option is20.- times_mad:
(Optional) For method
ratio_mad: Times the MAD to remove outliers. Default option is5.- times_zscore:
(Optional) For method
rna_counts_zscore: Times the zscore to remove outliers. Default option is3.
- sampling:
(Optional) Options for sampling counts and barcodes. Just for debug reasons.
- DNA:
Settings for sampling DNA counts.
- threshold:
Maximum threshold for DNA counts assigned to a BC.
- prop:
Sample down the DNA counts to this proporion.
- total:
Sample down the DNA counts to this number.
- seed:
Seed for the random DNA sampling.
- RNA:
Settings for sampling RNA counts.
- threshold:
Maximum threshold for RNA counts assigned to a BC.
- prop:
Sample down the RNA counts to this proporion.
- total:
Sample down the RNA counts to this number.
- seed:
Seed for the random RNA sampling.
Experiment file
Here we have 4 different options:
Forward, reverse, and UMI read
Experiment file has a header with Condition, Replicate, DNA_BC_F, DNA_UMI, DNA_BC_R, RNA_BC_F, RNA_UMI, and RNA_BC_R. Condition together with replicate have to be a uniqe name. Both field entries are not allowed to have _ and .. Multiple file names are allowd seperating them via ;. An example experiment file can be found here: resources/example_experiment.csv.
Condition,Replicate,DNA_BC_F,DNA_UMI,DNA_BC_R,RNA_BC_F,RNA_UMI,RNA_BC_R
HEPG2,1,SRR10800881_1.fastq.gz,SRR10800881_2.fastq.gz,SRR10800881_3.fastq.gz,SRR10800882_1.fastq.gz,SRR10800882_2.fastq.gz,SRR10800882_3.fastq.gz
HEPG2,2,SRR10800883_1.fastq.gz,SRR10800883_2.fastq.gz,SRR10800883_3.fastq.gz,SRR10800884_1.fastq.gz,SRR10800884_2.fastq.gz,SRR10800884_3.fastq.gz
HEPG2,3,SRR10800885_1.fastq.gz,SRR10800885_2.fastq.gz,SRR10800885_3.fastq.gz,SRR10800886_1.fastq.gz,SRR10800886_2.fastq.gz,SRR10800886_3.fastq.gz
Forward and reverse read
Experiment file has a header with Condition, Replicate, DNA_BC_F, DNA_BC_R, RNA_BC_F, and RNA_BC_R. Condition together with replicate have to be a uniqe name. Both field entries are not allowed to have _ and .. Multiple file names are allowd seperating them via ;.
Only forward read
Experiment file has a header with Condition, Replicate, DNA_BC_F, and RNA_BC_F. Condition together with replicate have to be a uniqe name. Both field entries are not allowed to have _ and .. Multiple file names are allowd seperating them via ;.
Forward, reverse, and UMI read using demultiplex option
Experiment file has a header with Condition, Replicate, BC_DNA, BC_RNA, BC_F, BC_R, UMI, and INDEX. Condition together with replicate have to be a uniqe name. Both field entries are not allowed to have _ and .. Multiple file names are allowd seperating them via ;.