Calculate Cosine distance between samples in an IDBac SQLite database

[chasemc]

R code to calculate Cosine distance between samples in an IDBac SQLite database:

install.packages(c("remotes", "dplyr", "DBI", "pool", "coop"))
remotes::install_github("chasemc/IDBacApp@*release")

library(IDBacApp)
library(dplyr)
library(DBI)
library(pool)
library(coop)

# Change this to be the full file path to your sqlite file 
sqlite_filepath <- tempfile(fileext = ".sqlite")

# don't run the download unless you are just testing 
download.file("https://massive.ucsd.edu/ProteoSAFe/DownloadResultFile?file=f.MSV000084291/raw/data/idbac_experiment_file.sqlite", destfile = sqlite_filepath)

db_pool <- IDBacApp::idbac_connect(fileName = tools::file_path_sans_ext(basename(sqlite_filepath)),
                                        filePath = dirname(sqlite_filepath))[[1]]
# Update the database to the latest version
idbac_update_db(pool = db_pool, copy_overwrite = "overwrite")

# Reconnect to the database
db_pool <- IDBacApp::idbac_connect(fileName = tools::file_path_sans_ext(basename(sqlite_filepath)),
                                   filePath = dirname(sqlite_filepath))[[1]]

# Peak at the tables in the SQLite database
sapply(DBI::dbListTables(conn = db_pool), 
       function(x){
            print(dplyr::tbl(db_pool, x ) %>%  head())
           }, simplify = F)


# Enter your sample IDs here
sample_ids <- c("172-1","172-10","172-11","172-7")

# Adjust the variables for mass cutoffs etc, below
temp <- lapply(sample_ids,
                   function(ids){
                       IDBacApp::idbac_get_peaks(pool = db_pool,
                                       sampleIDs = ids,
                                       minFrequency = 0,
                                       minNumber = NA, 
                                       lowerMassCutoff=3000,
                                       upperMassCutoff = 15000, 
                                       minSNR = 3, 
                                       tolerance = 0.002,
                                       type = "protein",   
                                       mergeReplicates = TRUE)[[1]]
                   })

names(temp) <- c(sample_ids)

# Adjust the variables for mass cutoffs etc, below. It's complicated but ppm isn't a direct cutoff, it controls how the peaks are "fuzzified"
fuzzy_vec <- IDBacApp::createFuzzyVector(
                    massStart =3000,
                    massEnd = 15000,
                    ppm = 1000,
                    massList = lapply(temp, function(x) x@mass),
                    intensityList = lapply(temp, function(x) x@intensity)
                    )

# Print distance matrix between samples
stats::as.dist(1 - coop::cosine(fuzzy_vec))