Shot Noise Analysis

This notebook analyzes the shot noise from TFA and MRFA analyses as discussed in Kantnerova et al. (2024)

Setup

Using R version 4.5.0 (2025-04-11) , tidyverse version 2.0.0, and isoorbi version 1.5.1.

# load packages
library(isoorbi) # for orbitrap function
library(dplyr) # for mutating data frames
library(forcats) # for recoding factors
library(ggplot2) # for data visualization
library(cowplot) # arrange multipanel plots

Trifluoroacetate (TFA)

Data from RAW (recommended)

# read files
raw_files <- 
  "data/shot_noise/tfa" |>
  orbi_find_raw() |>
  orbi_read_raw(include_spectra = c(1, 10, 100)) |>
  orbi_aggregate_raw() |>
  orbi_identify_isotopocules("data/shot_noise/tfa/isotopologs.tsv") |>
  orbi_filter_isotopocules()

✔ [109ms] orbi_read_raw() read TFA_1E4_AGC.raw from cache, included the spectra
from 3 scans

✔ [33ms] orbi_read_raw() read TFA_M0_1uscan_120kRes.raw from cache, included
the spectra from 3 scans

✔ [38ms] orbi_read_raw() read TFA_M0_1uscan_15kRes.raw from cache, included the
spectra from 3 scans

✔ [238ms] orbi_read_raw() finished reading 3 files

✔ [237ms] orbi_aggregate_raw() aggregated file_info (3), scans (17.51k), peaks
(91.64k), and spectra (2.08k) from 3 files using the standard aggregator

! [615ms] orbi_identify_isotopocules() identified 52.40k/91.64k peaks (57%)
representing 100% of the total ion current (TIC) as isotopocules M0, 13C, and
18O but encountered 2 warnings
  → ! isotopocule 18O matches multiple peaks in some same scans (5
  multi-matched peaks in total) - make sure to run orbi_flag_satellite_peaks()
  and orbi_plot_satellite_peak()
  → ! isotopocule 18O is missing from some scans (143 missing peaks in total) -
  make sure to evaluate coverage with e.g. orbi_plot_isotopocule_coverage()

✔ [7ms] orbi_filter_isotopocules() removed 39.38k / 91.79k peaks (43%) because
they were missing isotopocules (143), or unidentified peaks (39.24k). Remaining
isotopocules: M0, 13C, and 18O.

Bonus Figure: TFA Spectra

raw_files |> orbi_plot_spectra()

Figure 1: TFA spectra

Data from ISOX (legacy)

# originally, this is how the data was read in
isox_files <-
  "data/shot_noise/tfa" |>
  orbi_find_isox() |>
  orbi_read_isox() |>
  orbi_simplify_isox()

Shot noise calculation

# pre-process data
data_tfa <- raw_files |>
  orbi_flag_satellite_peaks() |>
  orbi_flag_weak_isotopocules(min_percent = 10) |>
  orbi_flag_outliers(agc_fold_cutoff = 2) |>
  orbi_define_basepeak("M0")

✔ [382ms] orbi_flag_satellite_peaks() flagged 5/52.40k peaks in 1 isotopocule
(18O) as satellite peaks (0.0095%)

✔ [13ms] orbi_flag_weak_isotopocules() confirmed there are no weak
isotopocules: all are detected in at least 10% of scans in each of the 9 data
groups (based on uidx, compound, and isotopocule)

✔ [9ms] orbi_flag_outliers() confirmed that none of the 17513 scans are
outliers based on 2 fold AGC cutoff, i.e. based on scans below 1/2 and above 2
times the average number of ions tic * it.ms in the Orbitrap analyzer, in 3
data groups (based on uidx)

✔ [661ms] orbi_define_basepeak() set M0 as the ratio denominator and calculated
34.89k ratio values for 2 isotopocules (13C and 18O)

# calculate shotnoise
shot_noise_tfa <- 
  data_tfa |>
  orbi_analyze_shot_noise() |>
  orbi_export_data_to_excel("output/shot_noise_tfa.xlsx")

✔ [13ms] orbi_analyze_shot_noise() analyzed the shot noise for 34.88k ratios
(excluding 5 flagged peaks)

✔ [2.7s] orbi_export_data_to_excel() exported the dataset (34.88k rows, 26
columns) to output/shot_noise_tfa.xlsx

Bonus Figure: TFA satellite peaks

data_tfa |> orbi_plot_satellite_peaks()

Figure 2: Satellite peaks

Bonus Figure: TFA isotopocule coverage

data_tfa |> orbi_plot_isotopocule_coverage()

Figure 3: Isotopocule coverage

Bonus Figure: TFA ratios

data_tfa |> orbi_plot_raw_data(y = ratio)

Figure 4: Isotopocule ratios vs. M0

Figure 10: TFA shotnoise vs counts/time

# individual plots
tfa_vs_ions <-
  shot_noise_tfa |>
  orbi_filter_isox(filename = "TFA_M0_1uscan_15kRes") |>
  orbi_plot_shot_noise(x = "n_effective_ions") +
  labs(title = "vs counts") +
  theme(plot.title = element_text(hjust = 0.5))

tfa_vs_time <-
  shot_noise_tfa |>
  orbi_filter_isox(filename = "TFA_M0_1uscan_15kRes") |>
  orbi_plot_shot_noise(permil_target = 1) +
  labs(title = "vs time") +
  theme(plot.title = element_text(hjust = 0.5))

# combine
plot_grid(
  tfa_vs_ions + theme(legend.position.inside = c(0.85, 0.74)), 
  tfa_vs_time + theme(legend.position = "none"),
  align = "h", nrow = 1, axis = "tb"
)

Figure 5: Shot noise vs counts and vs time.

Extended Data Figure 6: TFA shotnoise at different IT

shot_noise_tfa |>
  orbi_filter_isox(filename = c("TFA_1E4_AGC", "TFA_M0_1uscan_15kRes")) |>
  # change color legend
  mutate(
    IT_info = filename |>
      fct_recode(
        "IT = 0.03 ms" = "TFA_1E4_AGC",
        "IT = 0.75 ms" = "TFA_M0_1uscan_15kRes"
      )
  ) |>
  orbi_plot_shot_noise(color = "IT_info") +
  # wrap by the ratio label
  facet_wrap(~ratio_label) +
  theme(legend.position = c(0.41, 0.75))

Figure 6: Shot noise at different ITs.

Extended Data Figure 7: TFA shotnoise at different resolutions

shot_noise_tfa |>
  orbi_filter_isox(filename = c("TFA_M0_1uscan_120kRes", "TFA_M0_1uscan_15kRes")) |>
  # change color legend
  mutate(
    res_info = filename |>
      fct_recode(
        "res. = 120k" = "TFA_M0_1uscan_120kRes",
        "res. = 15k" = "TFA_M0_1uscan_15kRes"
      )
  ) |>
  orbi_plot_shot_noise(color = "res_info") +
  # wrap by the ratio label
  facet_wrap(~ratio_label) +
  theme(legend.position = c(0.42, 0.75))

Figure 7: Shot noise at different resolutions.

Model Peptide (MRFA)

Data from RAW (recommend)

# read files
raw_files_aas <- 
  "data/shot_noise/mrfa" |>
  orbi_find_raw() |>
  orbi_read_raw(include_spectra = c(1, 10, 100)) |>
  orbi_aggregate_raw() |>
  orbi_identify_isotopocules("data/shot_noise/mrfa/isotopologs.tsv") |>
  orbi_filter_isotopocules() |>
  orbi_calculate_ions()

✔ [250ms] orbi_read_raw() read MRFA_MS2_120kRes_5E5AGC_40NCE_10min_10uscans.raw
from cache, included the spectra from 3 scans

✔ [150ms] aggregate_files() aggregated file_info (1), scans (231), peaks
(158.31k), and spectra (33.11k) from 1 file using the standard aggregator

! [423ms] orbi_identify_isotopocules() identified 3.62k/158.31k peaks (2.3%) as
isotopcules M0, 15N, 13C, 33S, 2H, and 34S but encountered 2 warnings
  → ! isotopocule 13C matches multiple peaks in some same scans (31
  multi-matched peaks in total) - make sure to run orbi_flag_satellite_peaks()
  and orbi_plot_satellite_peak()
  → ! isotopocules 2H and 15N are missing from some scans (572 missing peaks in
  total) - make sure to evaluate coverage with e.g.
  orbi_plot_isotopocule_coverage()

✔ [3ms] orbi_filter_isotopocules() removed 155.26k / 158.88k peaks (98%)
because they were missing isotopocules (572), or unidentified peaks (154.69k).
Remaining isotopocules: M0, 15N, 13C, 33S, 2H, and 34S.

✔ [2ms] orbi_calculate_ions() calculated ions.incremental with noise factor CN
= 3 at reference resolution RN = 240000

Bonus Figure: MFRA Spectra

# Alanine mass window
raw_files_aas |> orbi_plot_spectra(mz_min = 43, mz_max = 46, max_scans = 1)

Figure 8: MRFA spectra - alanine

# Methionine mass window
raw_files_aas |> orbi_plot_spectra(mz_min = 102, mz_max = 110, max_scans = 1)

Figure 9: MRFA spectra - methionine

Data from ISOX (legacy)

# originally, this is how the data was read in
isox_files_aas <-
  "data/shot_noise/mrfa" |>
  orbi_find_isox() |>
  orbi_read_isox() |>
  orbi_simplify_isox()

Shot noise calculation

# load, process, and export mrfa data
data_aas <- 
  raw_files_aas |>
  orbi_flag_satellite_peaks() |>
  orbi_flag_weak_isotopocules(min_percent = 90) |> 
  orbi_flag_outliers(agc_fold_cutoff = 2) |> 
  orbi_define_basepeak("M0")

✔ [27ms] orbi_flag_satellite_peaks() flagged 31/3.62k peaks in 1 isotopocule
(13C) as satellite peaks (0.86%)

✔ [10ms] orbi_flag_weak_isotopocules() flagged 3 of 18 isotopocules as weak
because they were NOT present in at least 90% of scans in each of the 18 data
groups (based on uidx, compound, and isotopocule) → use
orbi_plot_isotopocule_coverage() to visualize them

✔ [6ms] orbi_flag_outliers() confirmed that none of the 231 scans are outliers
based on 2 fold AGC cutoff, i.e. based on scans below 1/2 and above 2 times the
average number of ions tic * it.ms in the Orbitrap analyzer

✔ [42ms] orbi_define_basepeak() set M0 as the ratio denominator and calculated
2.69k ratio values for 5 isotopocules (15N, 13C, 2H, 33S, and 34S)

shot_noise_aas <-
  data_aas |>
  orbi_analyze_shot_noise() |> 
  orbi_export_data_to_excel("output/shot_noise_MRFA.xlsx") 

✔ [10ms] orbi_analyze_shot_noise() analyzed the shot noise for 2.53k ratios
(excluding 165 flagged peaks)

✔ [259ms] orbi_export_data_to_excel() exported the dataset (2.53k rows, 26
columns) to output/shot_noise_MRFA.xlsx

Table

# example of first few rows
shot_noise_aas |>
  arrange(compound, isotopocule, scan.no) |>
  select(compound, scan.no, time.min, isotopocule,
         ratio, ratio_rel_se.permil, shot_noise.permil) |>
  head(10) |>
  knitr::kable()

compound	scan.no	time.min	isotopocule	ratio	ratio_rel_se.permil	shot_noise.permil
Alanine	1	0.0479792	15N	0.0038664	NaN	97.53305
Alanine	2	0.0916169	15N	0.0040361	21.48613	66.34978
Alanine	3	0.1352544	15N	0.0035002	41.68167	54.69087
Alanine	4	0.1789086	15N	0.0047781	66.84435	45.86767
Alanine	5	0.2225461	15N	0.0037879	53.97239	41.31846
Alanine	6	0.2661836	15N	0.0036734	46.67666	37.89584
Alanine	7	0.3098348	15N	0.0037989	39.96359	35.09461
Alanine	8	0.3534729	15N	0.0024645	61.49578	33.59419
Alanine	9	0.3971100	15N	0.0032222	57.26965	31.88088
Alanine	10	0.4407613	15N	0.0036329	51.23701	30.26539

Bonus Figure: Amino Acids satellite peaks

data_aas |> orbi_plot_satellite_peaks()

Figure 10: Satellite peaks

Bonus Figure: Amino Acids isotopocule coverage

data_aas |> orbi_plot_isotopocule_coverage()

Figure 11: Isotopocule coverage

Bonus Figure: Amino Acids ratios

data_aas |> orbi_plot_raw_data(y = ratio)

Figure 12: Isotopocule ratios vs. M0

Extended Data Figure 8: Amino Acids Shot Noise

shot_noise_aas |>
  filter(compound == "Methionine" | isotopocule != "2H") |>
  orbi_plot_shot_noise()

Figure 13: amino acids shotnoise