Package 'bppg'

Title: Bipartite peptide-protein graphs
Description: This package contains functionality to create and characterize bipartite graphs that model the relationship between peptides and proteins in bottom-up proteomics. With these graphs, protein ratios (fold change between two sample groups) are calculated from the respective measured peptide ratios. The main aim is to make use of quantitative information contained in shared peptides and making it possible to quantify proteins without shared peptides. Further, the package offers general analysis of peptide and protein relationships using bipartite graphs and visualization for these graphs.
Authors: Karin Schork [aut, cre] (ORCID: <https://orcid.org/0000-0003-3756-4347>), Julika Wenzel [aut] (ORCID: <https://orcid.org/0009-0004-1468-5984>), Katharina Neuhaus [aut], Martin Eisenacher [aut, fnd] (ORCID: <https://orcid.org/0000-0003-2687-7444>)
Maintainer: Karin Schork <[email protected]>
License: BSD_3_clause + file LICENSE
Version: 0.99.1
Built: 2026-10-11 10:05:26 UTC
Source: https://github.com/BiocStaging/bppg

Help Index

Adds vertex attributes with uniqueness of peptides and number of unique peptides for proteins.

Description

Adds vertex attributes with uniqueness of peptides and number of unique peptides for proteins.

Usage

.addUniquenessAttributes(G)

Arguments

G

igraph graph
Bipartrite peptide protein graph.

Value

A graph with 2 additional vertex attributes, uniqueness and nr_unique_peptides

See Also

generateGraphsFromFASTA(), generateQuantGraphs()

Examples

file <- system.file("extdata", "quantGraphs.rds", package = "bppg")
graphs <- readRDS(file)

res <- bppg:::.addUniquenessAttributes(graphs[[1]][[2]])

# get uniqueness of all nodes (NA means it is a protein node, not a peptide)
igraph::V(res)$uniqueness

# get number of unique and shared peptides per protein node:
igraph::V(res)$nr_unique_peptides
igraph::V(res)$nr_shared_peptides

Analyse results for a single protein

Description

Analyse results for a single protein

Usage

.analyseResultSingleProt(
  protNr,
  resProt,
  error_tol = 1e-10,
  ratioLog_tol = 1e-06
)

Arguments

protNr

integer(1)
Index of the protein to be analysed.
resProt

data.frame
The data.frame resulting from the iterateOverCi() function, filtered for a specific protein.
error_tol

numeric(1)
tolerance for the error term.
ratioLog_tol

numeric(1)
tolerance for the log protein ratios.

Value

Vector with minimal error, estimate for Ri (single value or min/max), estimate for Ci (single value or min/max).

Calculate equality and inequality constraints for the optimization step

Description

Calculate equality and inequality constraints for the optimization step

Usage

.calcConstraints(fixedCi, m)

Arguments

fixedCi

numeric vector
The fixed protein weights, variable weights set as NA. Sum of fixed weights must not exceed 1. If NULL, all Cis will be considered as variable. This argument is needed to fix Ci on a grid point in the iterated_Ci function.
m

integer(1)
number of proteins in the respective graph

Details

x is a vector containing first the RiLog values and then the Ci values (length 2*m if no Ci are fixed).

For the equality constraint (eqfun with corresponding bound eqB), the difference of the sum of the Ci values and 1 is calculated. The bound is set to 0, i.e. forcing the sum of the Ci to be 1. In case of fixed Ci values, the sum of the fixed and the free Ci values is considered.

The lower bound (LB) for the RiLog values is set to -Inf (no constraint) and to 0 for the Ci.

Value

list containing the following elements (see also solnp):

eqfun

function for calculating equality constraints
eqB

vector of equality bounds for the variables
LB

lower bound for the variables

Examples

# without fixed Ci
bppg:::.calcConstraints(fixedCi = NULL, m = 3)

# with fixed Ci
bppg:::.calcConstraints(fixedCi = c(NA, 0.8, NA), m = 3)

Calculate objective function for the optimization step

Description

Calculate objective function for the optimization step

Usage

.calcObjectiveFunction(fixedCi, M, rjLog)

Arguments

fixedCi

numeric vector
The fixed protein weights, variable weights set as NA. Sum of fixed weights must not exceed 1. If NULL, all Cis will be considered as variable. This argument is needed to fix Ci on a grid point in the iterated_Ci function.
M

matrix
Biadjacency matrix of the graph.
rjLog

numeric vector
Contains the measured peptide ratios (log2-scale).

Details

The objective function is the sum of squared error terms calculated in .errorEquation.

x is a vector containing first the RiLog values and then the Ci values (length 2*m if no Ci are fixed). If there are fixed Ci and m2 non-fixed ones, x has the length m + m2.

Value

objective function that will be minimized

Examples

M <- matrix(c(1, 1, 1, 0, 1, 1), ncol = 2, byrow = FALSE)
rjLog <- log2(c(0.6, 1.2, 1.5))

fun <- bppg:::.calcObjectiveFunction(fixedCi = NULL, M = M, rjLog = rjLog)

Calculate optimal solution for a single gridpoint

Description

Calculate optimal solution for a single gridpoint

Usage

.calcResultGridpoint(j, gridpoint, cnames, G, n, verbose, ...)

Arguments

j

integer(1)
index of protein
gridpoint

numeric(1)
Ci value to fix for this protein
cnames

character
column names for result dataframe
G

igraph object
bipartite peptide-protein graph
n

integer(1)
number of proteins in the graph
verbose

logical
If TRUE, print additional information (see solnp function).
...

additional arguments for .minimizeSquaredError(), e.g. verbose, control

Value

Vector with the following elements:

protein

index of protein (j)
grid

gridpoint
res_Ri_Ci

multiple values: estimated RiLog and Ci values
error

error term

Cleavage of a single protein sequence.

Description

Cleavage of a single protein sequence.

Usage

.cleave(sequence, start, stop, miss)

Arguments

sequence

character
The protein sequence.
start

integer vector
Index of where peptides start
stop

integer vector
Index of where peptides stop
miss

integer
which order of missed cleavages, e.g. 0 = no missed cleavage

Value

A dataframe with information (inkl. peptide sequence and start value) of the peptides.

See Also

.digest2(), digestFASTA()

Contracting of peptide and protein nodes.

Description

Contracting of peptide and protein nodes.

Usage

.contractGraph(G, vMapping, collProtNodes, collPeptNodes)

Arguments

G

igraph graph
Bipartrite peptide protein graph.
vMapping

list
A list with two dataframes from .getContractMapping().
collProtNodes

logical
If TRUE, the protein nodes will be collapsed.
collPeptNodes

logical
If TRUE, the peptide nodes will be collapsed.

Value

An edgelist with collapsed protein and/or peptide nodes.

See Also

For graph collapsing: .getContractMapping()
generateGraphsFromEdgelist()

Examples

library(seqinr)
file <- system.file("extdata", "uniprot_proteome_Scerevisiae_filtered.fasta",
     package = "bppg")
fasta <- seqinr::read.fasta(file = file, seqtype = "AA", as.string = TRUE)
edgelist <- digestFASTA(fasta)
vMapping <- bppg:::.getContractMapping(edgelist)
G <- igraph::graph_from_edgelist(as.matrix(edgelist), directed = FALSE)
igraph::V(G)[igraph::V(G)$name %in% edgelist[, 1]]$type <- TRUE
igraph::V(G)[igraph::V(G)$name %in% edgelist[, 2]]$type <- FALSE
res <- bppg:::.contractGraph(G, vMapping)

Digestion of a single protein sequence.

Description

Digestion of a single protein sequence.

Usage

.digest2(
  sequence,
  enzyme = "trypsin",
  missed = 0,
  warn = TRUE,
  remove_initial_M = TRUE
)

Arguments

sequence

character
The protein sequence.
enzyme

character
The enzyme used in digestion e.g. "trypsin" (does not cut before proline).
missed

character
The maximal number of missed cleavages.
warn

logical
If TRUE, warnings will be printed e.g. if a protein has no cleavage site.
remove_initial_M

logical
If TRUE, the initial methionine will be removed from the peptides.

Value

A vector of peptides.

See Also

digestFASTA()

Examples

library(seqinr)
file <- system.file("extdata", "uniprot_proteome_Scerevisiae_filtered.fasta",
    package = "bppg")
fasta <- seqinr::read.fasta(file = file, seqtype = "AA", as.string = TRUE)

# digest first protein sequence in the fasta file:
digested_proteins <- bppg:::.digest2(fasta[[1]])

Transform a bipartite graph into a directed graph.

Description

Transform a bipartite graph into a directed graph.

Usage

.directBipartiteGraph(bip_graph, from_type = FALSE)

Arguments

bip_graph

graph (igraph)
A bipartite graph.
from_type

logical
If TRUE, the edges will go out from the vertices with the type TRUE from the bipartite graph.

Value

A bipartite graph that is know directed.

Function to set up the error equations for the optimization problem

Description

Function to set up the error equations for the optimization problem

Usage

.errorEquation(RiLog, Ci, M, rjLog)

Arguments

RiLog

numeric vector
Contains the (estimated) protein ratios (log2-scale).
Ci

numeric vector
Contains the protein weights (estimated, sum up to 1)
M

matrix
The biadjaceny matrix of the corresponding graphs.
rjLog

numeric vector
Contains the measured peptide ratios (log2-scale).

Value

list containing the following elements:

res_Mat

matrix containing the estimated peptide ratios using the given Ri and Ci
res_equ

vector of error terms for each peptide
res_squ_err

sum of squared error terms
W

internal weight matrix

Examples

Ri <- c(0.5, 1.3)
Ci <- c(0.3, 0.7)
M <- matrix(c(1, 0, 1, 1), nrow = 2, byrow = TRUE)
rj <- c(0.6, 1.2)
bppg:::.errorEquation(Ri, Ci, M, rj)

Helper function that extracts the intensity columns and columns of interest from a given dataframe.

Description

Helper function that extracts the intensity columns and columns of interest from a given dataframe.

Usage

.extractIntensities(D, col_pattern, rename_columns)

Arguments

D

data.frame
Data frame of peptides.txt from MaxQuant
col_pattern

character
Pattern to recognize intensity columns by. Should be "Intensity." or "LFQ.intensity."
rename_columns

logical
If TRUE, "Intensity." or "LFQ.intensity." are removed

Value

returns intensity dataframe

Calculate peptide ratios for pairwise comparisons of groups (Y/X).

Description

Calculate peptide ratios for pairwise comparisons of groups (Y/X).

Usage

.foldChange(D, X, Y, useNA = FALSE)

Arguments

D

data.frame
The data set.
X

character
The column name of group1.
Y

character
The column name of group2.
useNA

logical
If TRUE, results 0 and Inf are possible, otherwise ratio is NA, if value for X or Y is NA

Value

The fold changes (Y/X).

Examples

D <- data.frame(s1 = c(1,4), s2 = c(2,5), s3 = c(3,6))
X <- "s1"
Y <- "s2"
bppg:::.foldChange(D,X,Y)

Calculate the geometric mean.

Description

Calculate the geometric mean.

Usage

.geomMean(x, useprod = FALSE, na.rm = FALSE)

Arguments

x

numeric vector
Input data.
useprod

logical
If TRUE, prod(x)^(1/n) will be calculated, otherwise exp(mean(log(x))).
na.rm

logical
If TRUE, missing values are removed before the calculation. FALSE is default.

Value

The geometric mean of the provided data points.

Examples

data <- c(1,6,3.5)
result <- bppg:::.geomMean(data, useprod = FALSE)

Create Mapping signature for igraph::contract function.

Description

Create Mapping signature for igraph::contract function.

Usage

.getContractMapping(edgelist, collProtNodes = TRUE, collPeptNodes = FALSE)

Arguments

edgelist

data.frame
An edgelist eg. created with digestFASTA().
collProtNodes

logical
If TRUE, the protein nodes will be collapsed.
collPeptNodes

logical
If TRUE, the peptide nodes will be collapsed.

Value

A list with two dataframes, one for peptide and one for protein signatures.

See Also

For collapsing graphs:
generateGraphsFromFASTA(), generateQuantGraphs()

Examples

library(seqinr)
file <- system.file("extdata", "uniprot_proteome_Scerevisiae_filtered.fasta",
     package = "bppg")
fasta <- seqinr::read.fasta(file = file, seqtype = "AA", as.string = TRUE)
edgelist <- digestFASTA(fasta)
res <- bppg:::.getContractMapping(edgelist)

Calulate initial values for Ci (protein weights) for optimization

Description

Calulate initial values for Ci (protein weights) for optimization

Usage

.initializeCi(fixedCi, m)

Arguments

fixedCi

numeric vector
The fixed protein weights, variable weights set as NA. Sum of fixed weights must not exceed 1. If NULL, all Cis will be considered as variable. This argument is needed to fix Ci on a grid point in the iterated_Ci function.
m

integer(1)
number of proteins in the respective graph

Details

If no Ci are fixed, all Ci are initialized with equal weights (1/m). If at least one Ci is fixed, the remaining weight (1 - sum of fixed Ci) is distributed equally among the non-fixed Ci.

Value

Ci_start: vector with initial start values for Ci for the optimization step

Examples

# without fixed Ci -> equal starting weights
bppg:::.initializeCi(fixedCi = NULL, m = 3)

# with fixed Ci -> fixes these and equal weight for the others
bppg:::.initializeCi(fixedCi = c(NA, 0.8, NA), m = 3)

Calulate initial values for Ri (protein ratios) for optimization

Description

Calulate initial values for Ri (protein ratios) for optimization

Usage

.initializeRi(M, rjLog)

Arguments

M

matrix
Biadjacency matrix of the graph.
rjLog

numeric
Vector of measured peptide ratios (log2-scale).

Details

For each protein, the mean of the peptide ratios (log2-scale) of the peptides belonging to this protein is calculated. If there are unique peptides for this protein, only those are used.

Value

RiLog_start: vector with inital start values for Ri for the optimization step (log2-scale)

Examples

M <- matrix(c(1, 1, 1, 0, 1, 1), ncol = 2, byrow = FALSE)
rjLog <- log2(c(0.6, 1.2, 1.5))
bppg:::.initializeRi(M, rjLog)

Enchanced version of the igraph::isomorphic function that also considers the node type in bipartite graphs, e.g. that W- and M-shaped graphs are NOT isomorphic

Description

Enchanced version of the igraph::isomorphic function that also considers the node type in bipartite graphs, e.g. that W- and M-shaped graphs are NOT isomorphic

Usage

.isomorphicBipartite(graph1, graph2, ...)

Arguments

graph1

graph (igraph)
First graph.
graph2

graph (igraph)
Second graph.
...

currently unused

Value

TRUE if graphs are isomorphic, FALSE if not.

See Also

generateGraphsFromEdgelist()

Examples

M1 <- matrix(c(1, 0, 1, 1), nrow = 2, byrow = TRUE)
G1 <- igraph::graph_from_biadjacency_matrix(M1)

M2 <- matrix(c(1, 1, 0, 1), nrow = 2, byrow = TRUE)
G2 <- igraph::graph_from_biadjacency_matrix(M2)

bppg:::.isomorphicBipartite(G1, G2)

Function to set up the optimization problem and minimize the sum of squared error terms

Description

Function to set up the optimization problem and minimize the sum of squared error terms

Usage

.minimizeSquaredError(G, fixedCi = NULL, verbose = FALSE, control = list())

Arguments

G

igraph object
bipartite peptide-protein graph
fixedCi

numeric vector
The fixed protein weights, variable weights set as NA. Sum of fixed weights must not exceed 1. If NULL, all Cis will be considered as variable. This argument is needed to fix Ci on a grid point in the iterated_Ci function.
verbose

logical
If TRUE, additional information on each iteration of the optimization is printed (see solnp).
control

list
The control parameters for solnp (see solnp).

Value

list containing the following elements:

RiLog

estimated protein ratios (Log2-scale)
Ci

estimated protein weights
RES

final result of .errorEquation, which also contains the final, minimal error term
Tracking

Tracking of Ri, Ci and error term for the different iterations (-1: no optimization)
outer.iter

Number of outer iterations needed for the optimization algorithm to converge or stop (see also solnp)
convergence

Indicates whether the solver has converged (0) or not (1 or 2) (see also solnp)

Examples

file <- system.file("extdata", "quantGraphs.rds", package = "bppg")
graphs <- readRDS(file)
G <- graphs$"1_2"[[2]]
bppg:::.minimizeSquaredError(G)

Set names for plotting with plotBipartiteGraph.

Description

Set names for plotting with plotBipartiteGraph.

Usage

.setNodeLabels(G, node_labels_peptides, node_labels_proteins, round_digits)

Arguments

G

igraph graph object
A bipartite peptide-protein graph.
node_labels_peptides

character
The type of labels for the peptides. Options are"numbers" or "pep_ratios" or "pep_ratio_aggr".
node_labels_proteins

character
The type of labels for the proteins. Options are "letters" or "acessions".
round_digits

integer
The number of digits to round the peptide ratios to.

Value

Graph with updated names

Funktion to format Tracking information into a dataframe with columnnames.

Description

Funktion to format Tracking information into a dataframe with columnnames.

Usage

.trackingDataFrame(i, RES, RiLog, Ci)

Arguments

i

numeric
iterrator, for when this tracking was performed
RES

list
The result from .errorEquation
RiLog

numeric vector
Contains the (estimated) protein ratios (log2-scale).
Ci

numeric vector
Contains the protein weights (estimated, sum up to 1)

Value

 \strong{data.frame} \cr
         combined information in one data.frame.

Aggregate replicates of the same experimental group.

Description

Aggregate replicates of the same experimental group.

Usage

aggregateReplicates(
  D,
  group = NULL,
  missing.limit = 0,
  method = "mean",
  seq_col = "Sequence"
)

Arguments

D

SummarizedExperiment
SummarizedExperiment Dataset containing peptide intensities, e.g. the result of normalizePeptideIntensities.
group

character factor
The groups per sample for aggregation, if not already in SummarizedExperiment::colData(D)$group.
missing.limit

numeric
The proportion of missing values that is allowed (e.g. 0 means no missings allowed).
method

character
The method of aggregation. Options are "mean", "sum" or "median"
seq_col

character
The column name containaining the peptide sequences in the rowData of the SummarizedExperiment. Default is "Sequence"

Value

SummarizedExperiment
A SummarizedExperiment with aggregated intensities ($intensities).

Examples

file <- system.file("extdata", "peptides_filtered.txt", package = "bppg")
group <- factor(rep(1:9, each = 3))
D <- readMqPeptideTable(path = file, group = group, LFQ = TRUE,
    remove_contaminants = FALSE)
D_norm <- bppg::normalizePeptideIntensities(D)
group <- factor(rep(1:9, each = 3))
aggregateReplicates(D_norm, group = group)

Extract protein ratio solutions from the result of iterateOverCi()

Description

Extract protein ratio solutions from the result of iterateOverCi()

Usage

automatedAnalysisIteratedCi(
  G,
  res,
  use_results_from_other_proteins = TRUE,
  verbose = FALSE,
  job = NULL,
  error_tol = 1e-10,
  ratioLog_tol = 1e-06
)

Arguments

G

igraph object
An igraph graph of the bipartite peptide-protein graph with peptide ratios
res

data.frame
The data.frame resulting from the iterateOverCi() function.
use_results_from_other_proteins

logical
If TRUE, the results from other proteins within the same graph will be used to calculate the optimal solution for each protein node. Default is TRUE
verbose

logical
If TRUE, additional information will be printed.
job

BatchExperiment job object
Is used to print the job id and parameters in the output
error_tol

numeric(1)
Tolerance for a constant error term. Default is 1e-10.
ratioLog_tol

numeric(1)
Tolerance for a constant log2 protein ratios. The default is 1e-6.

Value

A SummarizedExperiment object with one row for each protein in the assay.

See Also

iterateOverCi(), .minimizeSquaredError()

Examples

file <- system.file("extdata", "quantGraphs.rds", package = "bppg")
graphs <- readRDS(file)
G <- graphs$"1_2"[[2]]
# small example with a small grid size
res <- iterateOverCi(G, gridSize = 100)
automatedAnalysisIteratedCi(G, res)

Calculation of peptide ratios from aggregated intensities.

Description

Calculation of peptide ratios from aggregated intensities.

Usage

calculatePeptideRatios(D, group_levels = NULL)

Arguments

D

SummarizedExperiment
The result from function aggregateReplicates().
group_levels

character factor
The group information in the sample order. They are extracted from D by default

Value

A SummarizedExperiment with log2 peptide ratios (logRatios).

Examples

file <- system.file("extdata", "peptides_filtered.txt", package = "bppg")
group <- factor(rep(1:9, each = 3))
D <- readMqPeptideTable(path = file, group = group, LFQ = TRUE,
    remove_contaminants = FALSE)
D_norm <- bppg::normalizePeptideIntensities(D)
dAgg <- aggregateReplicates(D_norm)
calculatePeptideRatios(dAgg)

Combine relative quantification results over multiple pairwise comparisons into a single SummarizedExperiments object.

Description

Combine relative quantification results over multiple pairwise comparisons into a single SummarizedExperiments object.

Usage

combineComparisons(compResultList)

Arguments

compResultList

list
List of results to be concatenated. Each list element is the result of the optimization step for one pairwise comparison, i.e. a SummarizedExperiment object. Ideally, the element names should correspond to the comparison names.

Value

A SummarizedExperiment object with one assay per comparison. The row names of the assay matrices are the union of all row names across the input SummarizedExperiment objects. For rows that are not present in a given comparison, NA values are introduced. Row names are sorted alphabetically.

Examples

# prepared result list. Check "inst/scripts/filter_fasta_and_quantData.R" for
# details on how it was generated.
file <- system.file("extdata", "resultsList.rds", package = "bppg")
resultList <- readRDS(file)
SE <- combineComparisons(resultList)

In silico tryptic digestion of whole FASTA file.

Description

In silico tryptic digestion of whole FASTA file.

Usage

digestFASTA(
  fasta,
  missed_cleavages = 2,
  min_aa = 6,
  max_aa = 50,
  protOrigin = NULL,
  verbose = FALSE,
  ...
)

Arguments

fasta

list of character sequence
A fasta file, already read into R by seqinr::read.fasta(). If several protein origins are used they should be combined into one flatten list and a list with the corresponding origins should be provided for protOrigin.
missed_cleavages

integer
The maximal number of missed cleavages.
min_aa

integer
The minimal number of amino acids (set to 0 for no filtering).
max_aa

integer
The maximal number of amino acids (set to Inf for no filtering).
protOrigin

list
A list with the protein origin corresponding to fasta, proteins are used as index.
verbose

logical
If TRUE, additional information on each iteration of the digestion is printed.
...

Arguments passed on to .digest2

sequence

character
The protein sequence.

enzyme

character
The enzyme used in digestion e.g. "trypsin" (does not cut before proline).

missed

character
The maximal number of missed cleavages.

warn

logical
If TRUE, warnings will be printed e.g. if a protein has no cleavage site.

remove_initial_M

logical
If TRUE, the initial methionine will be removed from the peptides.

Value

data.frame with proteins and their peptide sequences, filtered for minimal and maximal number of amino acids.

See Also

.digest2()

Examples

library(seqinr)
file <- system.file("extdata", "uniprot_proteome_Scerevisiae_filtered.fasta",
    package = "bppg")
fasta <- seqinr::read.fasta(file = file, seqtype = "AA", as.string = TRUE)
res <- digestFASTA(fasta)

Export SummarizedExperiment object to an Excel file, with one assay per sheet.

Description

Export SummarizedExperiment object to an Excel file, with one assay per sheet.

Usage

exportSE(SE, file)

Arguments

SE

SummarizedExperiment object
Object to be exported. E.g. the result of combineComparisons.
file

character
file path to the output Excel file. If the file already exists, it will be overwritten.

Value

nothing, but an Excel file is written containing the assay matrices. Each assay is written to a separate sheet, and the sheet name corresponds to the assay name. Row names of the assay matrices are included in the Excel file.

Examples

# prepared result list. Check "inst/scripts/filter_fasta_and_quantData.R" for
# details on how it was generated.
file <- system.file("extdata", "resultsList.rds", package = "bppg")
resultList <- readRDS(file)
SE <- combineComparisons(resultList)
TMP <- tempdir()
exportSE(SE, file.path(TMP,"results.xlsx"))

Generate bipartite peptide-protein graphs from a list of digested proteins via an edgelist. Peptide and protein nodes can be contracted.

Description

Generate bipartite peptide-protein graphs from a list of digested proteins via an edgelist. Peptide and protein nodes can be contracted.

Usage

generateGraphsFromEdgelist(
  edgelist,
  collProtNodes = FALSE,
  collPeptNodes = FALSE
)

Arguments

edgelist

data.frame
An edgelist, output from digestFASTA(). Quant data needs to be in the column $pep_logRatio.
collProtNodes

logical
If TRUE, the protein nodes will be collapsed.
collPeptNodes

logical
If TRUE, the peptide nodes will be collapsed.

Value

A list of subgraphs as igraph objects.

See Also

digestFASTA()

Examples

library(seqinr)
file <- system.file("extdata", "uniprot_proteome_Scerevisiae_filtered.fasta",
     package = "bppg")
fasta <- seqinr::read.fasta(file = file, seqtype = "AA", as.string = TRUE)
edgelist <- digestFASTA(fasta)
res <- bppg::generateGraphsFromEdgelist(edgelist, collProtNodes = TRUE,
    collPeptNodes = TRUE)

Generate graphs from a FASTA file

Description

Generate graphs from a FASTA file

Usage

generateGraphsFromFASTA(
  fasta,
  collProtNodes = TRUE,
  collPeptNodes = TRUE,
  outpath = NULL,
  suffix = NULL,
  protOrigin = NULL,
  verbose = FALSE,
  ...
)

Arguments

fasta

list of vector of chars
A fasta file, already read into R by seqinr::read.fasta().
collProtNodes

logical
If TRUE, the protein nodes will be collapsed.
collPeptNodes

logical
If TRUE, the peptide nodes will be collapsed.
outpath

character
The path where intermediate results are saved. If NULL, results are not saved.
suffix

character
The suffix for saving results. will also be saved.
protOrigin

list or data.frame
A list with the protein orgin corresponding to fasta, proteins are used as rownames/index.
verbose

logical
If TRUE, additional information on each iteration of the optimization is printed
...

Additional arguments to bppg::digestFASTA()

Value

Subgraphs (i.e. connected components) from the graph generated from the FASTA file.

Examples

library(seqinr)
file <- system.file("extdata", "uniprot_proteome_Scerevisiae_filtered.fasta",
     package = "bppg")
fasta <- seqinr::read.fasta(file = file, seqtype = "AA", as.string = TRUE)
graphs <- bppg::generateGraphsFromFASTA(fasta)

Generate graphs from quantitative peptide-level data

Description

Generate graphs from quantitative peptide-level data

Usage

generateGraphsFromQuantData(
  D,
  fasta,
  outpath = NULL,
  missed_cleavages = 2,
  min_aa = 6,
  max_aa = 50,
  seq_column = "Sequence",
  norm_method = "loess",
  collProtNodes = TRUE,
  collPeptNodes = FALSE,
  suffix = "",
  protOrigin = NULL,
  verbose = FALSE,
  ...
)

Arguments

D

data.frame
A data set with peptide sequence as first column and peptide intensities in subsequent columns, e.g. created with readMqPeptideTable().
fasta

list of vector of characters
A fasta file used for identification of peptides in already read into R by seqinr::read.fasta().
outpath

character
The output path for the results.
missed_cleavages

integer
The number of allowed missed cleavages in a peptide.
min_aa

integer
The minimum number of amino acids in a peptide.
max_aa

integer
The maximum number of amino acids in a peptide.
seq_column

character
The column name of the column of D with the peptide sequences.
norm_method

character
The method of normalization. Options are "nonorm" (no normalization), "median", "loess", "quantile" or "lts" normalization. Default is "loess"
collProtNodes

logical
If TRUE, the protein nodes will be collapsed.
collPeptNodes

logical
If TRUE, the peptide nodes will be collapsed.
suffix

character
The suffix for output files.
protOrigin

list or data.frame
A list with the protein orgin corresponding to fasta, proteins are used as rownames/index.
verbose

logical
If TRUE, additional information on each iteration of the optimization is printed
...

Additional arguments for .digest2().

Value

A list of list of graphs. The outer list is for the different possible comparisons, the inner layer is for the independet graphs in that comparison.

See Also

readMqPeptideTable(), seqinr::read.fasta(), generateQuantGraphs(), generateGraphsFromFASTA()

Examples

library(seqinr)
file <- system.file("extdata", "uniprot_proteome_Scerevisiae_filtered.fasta",
    package = "bppg")
fasta <- seqinr::read.fasta(file = file, seqtype = "AA", as.string = TRUE)

file <- system.file("extdata", "peptides_filtered.txt", package = "bppg")
group <- factor(rep(1:9, each = 3))
D <- readMqPeptideTable(path = file, group = group, LFQ = TRUE,
    remove_contaminants = FALSE)

graphs <- bppg::generateGraphsFromQuantData(D, fasta)

Generate graphs from peptide ratio table, using an edgelist calculated on the fasta file.

Description

Generate graphs from peptide ratio table, using an edgelist calculated on the fasta file.

Usage

generateQuantGraphs(
  exp_peptide_ratios,
  fasta_edgelist,
  seq_column = "Sequence",
  outpath = NULL,
  collProtNodes = TRUE,
  collPeptNodes = FALSE,
  suffix = ""
)

Arguments

exp_peptide_ratios

SummarizedExperiment
A SummarizedExperiment from calculatePeptideRatios with peptide ratios.
fasta_edgelist

data.frame
An edgelist created from the corresponding FASTA file, eg. created with digestFASTA().
seq_column

character
The column name of the peptide sequence.
outpath

character
The output path for the results.
collProtNodes

logical
If TRUE, the protein nodes will be collapsed.
collPeptNodes

logical
If TRUE, the peptide nodes will be collapsed.
suffix

character
The suffix for saving results.

Value

A list of list of subgraphs

See Also

digestFASTA()

Examples

library(seqinr)
file <- system.file("extdata", "uniprot_proteome_Scerevisiae_filtered.fasta",
     package = "bppg")
fasta <- seqinr::read.fasta(file = file, seqtype = "AA", as.string = TRUE)
edgelist <- digestFASTA(fasta)

file <- system.file("extdata", "peptides_filtered.txt", package = "bppg")
group <- factor(rep(1:9, each = 3))
D <- readMqPeptideTable(path = file, group = group, LFQ = TRUE,
     remove_contaminants = FALSE)
D_norm <- bppg::normalizePeptideIntensities(D)
dAgg <- aggregateReplicates(D_norm)
exp_peptide_ratios <- calculatePeptideRatios(dAgg)

res <- generateQuantGraphs(exp_peptide_ratios, edgelist)

Iterate over possible Ci values

Description

Iterate over possible Ci values

Usage

iterateOverCi(
  G,
  gridStart = 0,
  gridStop = 1,
  gridSize = 1000,
  omit_grid_borders = TRUE,
  extend_grid_at_borders = FALSE,
  verbose = FALSE,
  control = list()
)

Arguments

G

igraph object
Bipartite peptide-protein graph
gridStart

integer
The start of the grid (default is 0).
gridStop

integer
The end of the grid (default is 1).
gridSize

integer
The number of grid points for the Cis.
omit_grid_borders

logical
If TRUE (default), omit exact value of 1 and 0 from the grid (recommended, as they may cause numerical issues).
extend_grid_at_borders

logical
If TRUE, the grid will be extend close to the borders (0 and 1). While exact values of 0 and 1 may cause numerical problems, values close to those may be valuable to get a better estimate of the protein ratios. Default is FALSE.
verbose

logical
If TRUE, print additional information (see solnp function).
control

list
The control object to be passed to the .minimizeSquaredError() function.

Details

With .minimizeSquaredError() each protein node receives one estimate for the protein ratio. However, in some cases, there are multiple possible values for the protein ratios that lead to the same, minimal error term. To get a better coverage of the optimal solutions, this function iterates over a grid of possible weights Ci for each protein node. For each protein node, the Ci value if fixed on a point on the grid, while the other Ci values and all Ri values are optimized using .minimizeSquaredError(). The resulting table can be used to assess a range of possible solutions for the protein ratios. The table can be further processed with automatedAnalysisIteratedCi().

Value

A data.frame containing the optimal Ci and Ri values together with the reached minimal error term for each grid point.

See Also

.minimizeSquaredError(), automatedAnalysisIteratedCi()

Examples

file <- system.file("extdata", "quantGraphs.rds", package = "bppg")
graphs <- readRDS(file)
G <- graphs$"1_2"[[2]]
# small example with a small grid size
iterateOverCi(G, gridSize = 100)

Normalization of peptide intensities

Description

Normalization of peptide intensities

Usage

normalizePeptideIntensities(
  D,
  method = "loess",
  lts.quantile = 0.8,
  log_base = 2
)

Arguments

D

SummarizedExperiment
SummarizedExperiment dataset containing peptide intensities, e.g. readMqPeptideTable.
method

character
The method of normalization. Options are "nonorm" (no normalization), "median", "loess", "quantile" or "lts" normalization. Default is "loess"
lts.quantile

numeric
The quantile for the lts normalization. Default is 0.8.
log_base

numeric
The base for log-transformation. Default is 2.

Value

SummarizedExperiment
Dataset containing normalized peptide intensities.

Examples

file <- system.file("extdata", "peptides_filtered.txt", package = "bppg")
D <- readMqPeptideTable(path = file, LFQ = TRUE, remove_contaminants = FALSE)
D_norm <- normalizePeptideIntensities(D, method = "loess")

Plotting of bipartite peptide-protein graphs.

Description

Plotting of bipartite peptide-protein graphs.

Usage

plotBipartiteGraph(
  G,
  legend = TRUE,
  vertex.color = c("mediumseagreen", "cadetblue2", "coral1"),
  vertex.size = 15,
  edge.width = 0.5,
  node_labels_proteins = "letters",
  node_labels_peptides = "numbers",
  round_digits = 2,
  output_path = NULL,
  ...
)

Arguments

G

igraph graph object
A bipartite peptide-protein graph.
legend

logical
If TRUE, a legend will be added.
vertex.color

character vector
The colours for the different vertex types.
vertex.size

numeric
The size of vertices.
edge.width

numeric
The width of the edges.
node_labels_proteins

character
The type of labels for the proteins. Options are "letters" or "accessions".
node_labels_peptides

character
The type of labels for the peptides. Options are "numbers" or "pep_logRatios" or "pep_logRatios_mean".
round_digits

integer
The number of digits to round the peptide ratios to.
output_path

character
file path for optional save of figure.
...

Additional arguments for ggsave.

Value

Plot of one bipartite graph.

Examples

file <- system.file("extdata", "quantGraphs_collpept.rds", package = "bppg")
graphs <- readRDS(file)
G <- graphs$"1_2"[[2]]

plotBipartiteGraph(G)

Tries to recursively remove protein nodes from graph while keeping the error term in the optimization step low

Description

Tries to recursively remove protein nodes from graph while keeping the error term in the optimization step low

Usage

proteinElimination(
  G,
  threshold = 1.05,
  control = list(),
  min_error_ref = NULL,
  resDF = NULL,
  protsOriginIDs = NULL,
  res_best = NULL
)

Arguments

G

igraph graph object
The graph to eliminate the proteins in.
threshold

numeric
The threshold for increase of error term. The default 1.05 refers to 5% increase.
control

list
A list of control parameters for the optimization step. See solnp for details.
min_error_ref

numeric
The minimal error term on the whole graph using all available protein nodes.
resDF

data.frame
Dataframe for collecting results from each iteration. It contains information on the tested combinations of protein nodes, the error terms and the currently best combination.
protsOriginIDs

igraph node list
This will be initialized during the first iteration, please keep the default NULL value. It is the current list of protein nodes relative to the original graph.
res_best

list
A list containing the currently best solution including the protein node combination, the error term and the corresponding graph object.

Details

This function works in a recursive way. In the first iteration, the error term on the whole graph is assessed. Then, one of the protein nodes is deleted and the function is recursively applied. The results of all iterations are collected in the resDF data.frame.

For starting the first iteration, only G, threshold and if necessary control have to be defined, everything else will be calculated during the first iteration for all future iterations.

Value

list containing the following elements:

min_error_ref

reference error term of the whole graph
protsOriginIDs

list of all available protein nodes in the beginning
resDF

dataframe with results of all iterations
res_best

List of the overall best solution. The graph object(s) in this list can directly be used by iterateOverCi.

See Also

.minimizeSquaredError()

Examples

## TODO
file <- system.file("extdata", "quantGraphs.rds", package = "bppg")
graphs <- readRDS(file)
G <- graphs[[1]][[4]]
# This graph has 4 protein nodes.

proteinElimination(G)
# Only 3 protein nodes are kept after protein elimination.
# The error term only increases slightly.

Import of FragPipe's combined_peptide.tsv-table.

Description

Import of FragPipe's combined_peptide.tsv-table.

Usage

readFpPeptideTable(
  path,
  group = NULL,
  rename_columns = TRUE,
  zeroToNA = TRUE,
  remove_empty_rows = TRUE,
  further_columns_to_keep = NULL,
  verbose = FALSE
)

Arguments

path

character
The path to the peptides.txt table.
group

character
List or vector of group names corresponding to the order of samples.
rename_columns

logical
If TRUE, "Intensity." or "LFQ.intensity." are removed
zeroToNA

logical
If TRUE, zeros are converted to NAs.
remove_empty_rows

logical
If TRUE, rows with only NAs are removed.
further_columns_to_keep

integer vector
Indices of additional columns to keep, except peptide sequence and intensities
verbose

logical
If TRUE, additional information on the data extraction is printed

Value

A SummarizedExperiment with intensities, sequences, and optional data for the rowData dataframe.

Examples

file <- system.file("extdata", "peptides_FP_filtered.tsv", package = "bppg")
D1 <- bppg::readFpPeptideTable(file)

Import of MaxQuant's peptide.txt-table.

Description

Import of MaxQuant's peptide.txt-table.

Usage

readMqPeptideTable(
  path,
  group = NULL,
  LFQ = FALSE,
  remove_contaminants = FALSE,
  rename_columns = TRUE,
  zeroToNA = TRUE,
  remove_empty_rows = TRUE,
  further_columns_to_keep = NULL,
  verbose = FALSE
)

Arguments

path

character
The path to the peptides.txt table.
group

character
List or vector of group names corresponding to the order of samples.
LFQ

logical
If TRUE, LFQ intensities are used, if FALSE, raw (unnormalized) intensities
remove_contaminants

logical
If TRUE, peptide sequences from potential contaminants are removed
rename_columns

logical
If TRUE, "Intensity." or "LFQ.intensity." are removed
zeroToNA

logical
If TRUE, zeros are converted to NAs.
remove_empty_rows

logical
If TRUE, rows with only NAs are removed.
further_columns_to_keep

integer vector
Indices of additional columns to keep, except peptide sequence and intensities
verbose

logical
If TRUE, additional information on the data extraction is printed

Value

A SummarizedExperiment with intensities, sequences, and optional data for the rowData dataframe.

Examples

file <- system.file("extdata", "peptides_filtered.txt", package = "bppg")
D <- readMqPeptideTable(path = file, LFQ = TRUE, remove_contaminants = FALSE)