Model and analyze MS/MS-based protein melting data.

Description

mstherm is a package for modeling and analysis of MS/MS-based thermal proteome profiling (TPP) experiments.

Author(s)

Jeremy Volkening jdv@base2bio.com

Create a new MSThermExperiment.

Description

MSThermExperiment creates a new experiment object from a set of filenames or data frames.

Usage

MSThermExperiment(control, annotations)

Arguments

Details

Both parameters can take either a data frame or a tab-delimited filename on disk (which will be read into a data frame). "control" should contain columns with the following headers (in any order):

"name"

Unique identifier of a single replicate

"sample"

Sample name that a replicate belongs to

"data_file"

Path to file on disk containing the quantification data

"meta_file"

Path to file on disk containing the labeling metadata

The "meta_file" should be tab-delimited text and contain two columns labeled "channel" and "temp". The "data_file" should be tab-delimited text and contain, at a minimum, the following columns:

"peptide"

Sequence of the matched peptide in single-letter IUPAC

"protein"

Protein or protein group to which the peptide belongs

"..."

One column per isobaric channel, containing absolute quantification values. Column names must match those in the "channel" column of the meta file, with the exception that R will automatically convert any name not compatible with its syntax rules. To be safe, use only letters, digits, underscores, and periods in channel names and never start with a digit (e.g. use "TMT.126" rather than "126")

The following columns can also be utilized for filtering if included (all others will simply be ignored):

"coelute_inf"

Calculated precursor co-isolation interference (0.0-1.0)

"score"

Score assigned by the processing software to the PSM

"annotations" should contain two columns with the headers "name" and "annotation". "name" should match the protein names in the data files, and "annotation" can contain any text (generally a functional description)

Value

An MSThermExperiment object

Examples

control <- system.file("extdata", "demo_project/control.tsv", package="mstherm")
annots  <- system.file("extdata", "demo_project/annots.tsv",  package="mstherm")
expt    <- MSThermExperiment(control, annotations=annots)

Convert absolute quantitation to relative ratios.

Description

abs_to_ratio takes a vector of absolute values and returns a vector of ratios relative to some starting point.

Usage

abs_to_ratio(x, method = "first")

Arguments

Details

The denominator used to calculate relative protein concentrations can affect the ability to model noisy data. In the theoretically ideal scenario, everything would be relative to the lowest temperature point. However, other methods can be used to help alleviate problems related to noise. Available methods include:

"first"

Use the first value (lowest temperature point) (default)

"max"

Use the maximum value

"top3"

Use the mean of the three highest values

"near"

Use the median of all values greater than 80 the first value

Value

A numeric vector of the same length as input

MSResultSet to data frame.

Description

Populates a data frame with information from an MSResultSet, one row per protein/group

Usage

## S3 method for class 'MSThermResultSet'
as.data.frame(x, ...)

Arguments

Value

A data frame populated with relevant information per result

Examples

control <- system.file("extdata", "demo_project/control.tsv", package="mstherm")
annots  <- system.file("extdata", "demo_project/annots.tsv",  package="mstherm")
expt    <- MSThermExperiment(control, annotations=annots)
expt    <- normalize_to_std(expt, "cRAP_ALBU_BOVIN", plot=FALSE)
res     <- model_experiment(expt, bootstrap=FALSE, np=2)

df <- as.data.frame(res)

Generate protein ratio profile from spectrum quantification matrix.

Description

gen_profile takes a matrix of spectrum channel quantification values belonging to a protein and "rolls them up" into a vector of protein-level relative quantification values.

Usage

gen_profile(x, method = "sum", method.denom = "first")

Arguments

Details

The following methods for spectrum-to-protein conversion are supported:

"sum"

use the sum of the Spectrum values for each channel

"median"

use the median of the spectrum values for each channel

"ratio.median"

Like "median", but values for each spectrum are first converted to ratios according to "method.denom" channel

"ratio.mean"

Like "ratio.median" but using mean of ratios

Value

A numeric vector of the same length as the number of matrix columns

Model MSThermExperiment.

Description

Model multiple proteins from an MSThermExperiment object.

Usage

model_experiment(expt, proteins, np, ...)

Arguments

Value

MSThermResultSet object

Examples

control <- system.file("extdata", "demo_project/control.tsv", package="mstherm")
annots  <- system.file("extdata", "demo_project/annots.tsv",  package="mstherm")
expt    <- MSThermExperiment(control, annotations=annots)
expt    <- normalize_to_std(expt, "cRAP_ALBU_BOVIN", plot=FALSE)

res     <- model_experiment(expt, bootstrap=FALSE, np=2)
summary(res)

Model single protein.

Description

Model a single protein from an MSThermExperiment object.

Usage

model_protein(expt, protein, min_rep_psm = 0, min_smp_psm = 0,
  min_tot_psm = 0, max_inf = 1, min_score, max_score, smooth = 0,
  method = "sum", method.denom = "near", trim = 0, bootstrap = 0,
  min_bs_psms = 8, annot_sep = "|", max_slope = 0, min_r2 = 0,
  min_reps = 0, only_modeled = 0, check_missing = 0,
  missing_cutoff = 0.3)

Arguments

Details

Valid quantification methods include:

"sum"

use the sum of the spectrum values for each channel

"median"

use the median of the spectrum values for each channel

"ratio.median"

Like "median", but values for each spectrum are first converted to ratios according to "method.denom" channel

"ratio.mean"

Like "ratio.median" but using mean of ratios

Valid denominator methods include:

"first"

Use the first value (lowest temperature point) (default)

"max"

Use the maximum value

"top3"

Use the mean of the three highest values

"near"

Use the median of all values greater than 80 the first value

Value

MSThermResult object

Examples

control <- system.file("extdata", "demo_project/control.tsv", package="mstherm")
annots  <- system.file("extdata", "demo_project/annots.tsv",  package="mstherm")
expt    <- MSThermExperiment(control, annotations=annots)
expt    <- normalize_to_std(expt, "cRAP_ALBU_BOVIN", plot=FALSE)

model   <- model_protein(expt, "P38707", smooth=TRUE, bootstrap=FALSE)
summary(model)

Normalize to a profile.

Description

Normalizes an MSThermReplicate based on a pre-determined vector of relative abundances

Usage

normalize_to_profile(replicate, profile, model = T, plot = T)

Arguments

Value

An MsThermReplicate object with normalized data slots

Examples

control <- system.file("extdata", "demo_project/control.tsv", package="mstherm")
annots  <- system.file("extdata", "demo_project/annots.tsv",  package="mstherm")
expt    <- MSThermExperiment(control, annotations=annots)

profile <- c(50.0, 50.5, 47.5, 42.0, 37.0, 25.0, 16.0, 11.5, 10.5, 10.0)
expt$samples$Control$replicates$C1 <- normalize_to_profile(
   expt$samples$Control$replicates$C1, profile, plot=FALSE
)

Normalize to a spike-in standard.

Description

Normalizes each replicate of an experiment based on a given spike-in protein standard (assumed to be present in equimolar amounts in each channel).

Usage

normalize_to_std(expt, protein, model = T, plot = T)

Arguments

Value

An MsThermExperiment object with normalized data slots

Examples

control <- system.file("extdata", "demo_project/control.tsv", package="mstherm")
annots  <- system.file("extdata", "demo_project/annots.tsv",  package="mstherm")
expt    <- MSThermExperiment(control, annotations=annots)

expt    <- normalize_to_std(expt, "cRAP_ALBU_BOVIN", plot=FALSE)

Re-normalize based on Tm.

Description

Normalizes each replicate of an experiment based on linear regression of calculated Tm (corrects for remaining systematic error).

Usage

normalize_to_tm(expt, res)

Arguments

Details

An assumption can be made in most TPP experiments using a single organism that the Tm of most proteins should not be changing. However, global shifts have been observed between replicates, presumably due to technical variance, which complicate data interpretation. This method attempts to remove this source of error by doing a bootstrap renormalization of the quantification values based on pairwise linear regression between calculated Tms of each replicate. A reference set of Tms is calculated based on all replicates, and each replicate is normalized to this based on the calculated slope and intercept of the input data.

Value

An MsThermExperiment object with re-normalized data slots

Examples

control <- system.file("extdata", "demo_project/control.tsv", package="mstherm")
annots  <- system.file("extdata", "demo_project/annots.tsv",  package="mstherm")
expt    <- MSThermExperiment(control, annotations=annots)
expt    <- normalize_to_std(expt, "cRAP_ALBU_BOVIN", plot=FALSE)
res     <- model_experiment(expt, smooth=TRUE, bootstrap=FALSE, np=2)

expt    <- normalize_to_tm(expt, res)

Plot MSThermResult object.

Description

Generate a denaturation plot for an modeled protein/group.

Usage

## S3 method for class 'MSThermResult'
plot(x, table = T, col, CI.points = T, CI.Tm = T,
  ...)

Arguments

Value

Nothing

Examples

control <- system.file("extdata", "demo_project/control.tsv", package="mstherm")
annots  <- system.file("extdata", "demo_project/annots.tsv",  package="mstherm")
expt    <- MSThermExperiment(control, annotations=annots)
expt    <- normalize_to_std(expt, "cRAP_ALBU_BOVIN", plot=FALSE)
res     <- model_experiment(expt, bootstrap=FALSE, np=2)

# plot single MSThermResult
plot(res$P38707)

# plot all proteins (e.g. to pdf device, one-per-page)
plot(res)

Plot MSThermResultSet object.

Description

Generate a series of denaturation plots for all results in an MSThermResultSet.

Usage

## S3 method for class 'MSThermResultSet'
plot(x, ...)

Arguments

Details

Since this function makes multiple sequential calls to plot.MSThermResult, it is usually used in conjunction with a multipage graphics device such as "pdf()". Otherwise each subsequent call will only overwrite the previous output.

Value

Nothing

Examples

# see plot.MSThermResult for an example

Summarize MSThermResult object.

Description

Print a summary of an MSThermResult, including samples and parameters.

Usage

## S3 method for class 'MSThermResult'
summary(object, ...)

Arguments

Value

Nothing

Examples

# see model_protein() for an example

Summarize MSThermResultSet object.

Description

Print a summary of an MSThermResultSet, including samples and parameters.

Usage

## S3 method for class 'MSThermResultSet'
summary(object, ...)

Arguments

Value

Nothing

Examples

# see model_experiment() for an example

Export MSThermResultSet to an SQLite database.

Description

Exports and MSThermResultSet object to a new SQLite database file. Each model (specific to a given replicate and protein) is exported as an individual record. The schema used for the 'data' table can be seen in the code below.

Usage

write.sqlite(res, file)

Arguments

Value

Nothing

Examples

control <- system.file("extdata", "demo_project/control.tsv", package="mstherm")
annots  <- system.file("extdata", "demo_project/annots.tsv",  package="mstherm")
expt    <- MSThermExperiment(control, annotations=annots)
expt    <- normalize_to_std(expt, "cRAP_ALBU_BOVIN", plot=FALSE)
res     <- model_experiment(expt, bootstrap=FALSE, np=2)

fn <- tempfile(fileext = ".sqlite")
write.sqlite(res, fn)
unlink(fn) # for example only