In jaspershen/lipidflow: What the Package Does (One Line, Title Case)

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  out.width = "100%"
)

---

This step is used to get the relative quantification data of internal standards and lipids in all samples. If you have not finished the previous step, please click here:

Get retention time of internal standards.

If you have finished this step, click here to next step:

Absolute quantification.

---

Get relative quantification

We should have finished step 1. Then we combine the internal standard information from positive and negative mode.

---

Combine internal standard information

library(lipidflow)
library(tidyverse)
library(openxlsx)

is_info_table_new_pos = 
  readxl::read_xlsx("example/POS/IS_info_new.xlsx")

is_info_table_new_neg = 
  readxl::read_xlsx("example/NEG/IS_info_new.xlsx")

is_info_table_new =
  is_info_table_new_pos %>%
  dplyr::left_join(is_info_table_new_neg[, c("name", "rt_neg_second", "rt_neg_min", "adduct_neg", "mz_neg")],
                   by = "name")

head(is_info_table_new)

Then we output the is_info_table_new into the example/Result folder.

We first create a folder named as Result:

dir.create("example/Result")

dir.create("example/Result")
openxlsx::write.xlsx(is_info_table_new,
                     file = "example/Result/IS_info_table.xlsx")

---

Relative quantification

1. Positive mode

We need to get the sample information of samples first.

sample_info_pos =
  generate_sample_info(path = "example/POS")

head(sample_info_pos)

Get the Cholesterol retention time:

if (any(is_info_table_new$name == "Cholesterol")) {
  idx = which(is_info_table_new$name == "Cholesterol")
  chol_rt2 = c(is_info_table_new$rt_pos_second[idx],
               is_info_table_new$rt_neg_second[idx])
  chol_rt2 = chol_rt2[!is.na(chol_rt2)]
  if (length(chol_rt2) > 0) {
    chol_rt = chol_rt2[1]
  }
} else{
  chol_rt = 1169
}

chol_rt

Internal standards

Then run get_relative_quantification() to get the relative quantification data of internal standards:

get_relative_quantification(
  path = "example/POS",
  output_path_name = "is_relative_quantification",
  targeted_table_name = "IS_info_new.xlsx",
  sample_info = sample_info_pos,
  targeted_table_type = "is",
  polarity = "positive",
  chol_rt = chol_rt,
  output_eic = TRUE,
  ppm = 40,
  rt.tolerance = 180,
  threads = 5,
  rerun = FALSE
)

The meanings of parameters:

1. path: Work directory. Here we set as example/POS.

2. targeted_table_name: The name of internal standard information.

3. sample_info: sample information of samples.

4. targeted_table_type: The type of extracted peaks, "is" (internal standard) or "lipid" (lipids).

5. polarity: "positive" or negative.

6. chol_rt: Cholesterol retention time (second).

7. output_eic: Output EICs of peaks or not.

8. ppm: peak detection ppm.

9. rt.tolerance: peak detection retention time tolerance (second).

10. threads: Number of cores to run.

11. rerun: Rerun peak detection or not.

Then all the results are outputted into a folder named as example/POS/is_relative_quantification. The relative quantification table is is_quantification_table.xlsx and all the peak shapes are in peak_shape.

is_quantification_table.xlsx is:

One example of peak shape (15_0-18_1(d7) PC):

Rectangle region means the integrate of peak region.

htmltools::tags$iframe(
  title = "Internal standard",
  src = "../man/figures/15_0_18_1_d7_PC.html",
  width = "100%",
  height = "600",
  scrolling = "no",
  seamless = "seamless",
  frameBorder = "0"
)

Lipids

Then run get_relative_quantification() for lipids:

get_relative_quantification(
  path = "example/POS",
  output_path_name = "lipid_relative_quantification",
  targeted_table_name = "lipid_annotation_table_pos.xlsx",
  sample_info = sample_info_pos,
  targeted_table_type = "lipid",
  polarity = "positive",
  chol_rt = chol_rt,
  output_eic = TRUE,
  ppm = 40,
  rt.tolerance = 180,
  threads = 5,
  rerun = FALSE
)

Here the targeted_table_name is the annotation table from lipidsearch. It should be placed in POS and NEG folders, respectively.

The results are outputted in example/POS/lipid_relative_quantification folder. These results are similar with internal standards.

2. Negative mode

Negative mode is same with positive mode.

sample_info_neg =
  generate_sample_info(path = "example/NEG")

Internal standards

get_relative_quantification(
  path = "example/NEG",
  output_path_name = "is_relative_quantification",
  targeted_table_name = "IS_info_new.xlsx",
  sample_info = sample_info_neg,
  targeted_table_type = "is",
  polarity = "negative",
  chol_rt = chol_rt,
  output_integrate = TRUE,
  output_eic = TRUE,
  ppm = 40,
  rt.tolerance = 180,
  threads = 5,
  rerun = FALSE
)

Lipids

Then run get_relative_quantification() for lipids:

get_relative_quantification(
  path = "example/NEG",
  output_path_name = "lipid_relative_quantification",
  targeted_table_name = "lipid_annotation_table_neg.xlsx",
  sample_info = sample_info_neg,
  targeted_table_type = "lipid",
  polarity = "negative",
  chol_rt = chol_rt,
  output_eic = TRUE,
  ppm = 40,
  rt.tolerance = 180,
  threads = 5,
  rerun = FALSE
)

---

Manually check for relative quantification data of internal standards

Sometimes, the peak detection maybe not accurate, so we need to manually check the peak shape of all the internal standards and then detect peak and integrate peak shape again.

For example, the internal standard 18_1_d7_ Lyso PC in positive mode (example/POS/is_relative_quantification/peak_shape/18_1_d7_ Lyso PC.html), the peak shape is like below figure shows:

htmltools::tags$iframe(
  title = "Internal standard",
  src = "../man/figures/figure31.html",
  width = "100%",
  height = "600",
  scrolling = "no",
  seamless = "seamless",
  frameBorder = "0"
)

From this figure, we can see that for the samples D25_1, D25_2 and M19_2, the integrate region (begin point) is not correct, so we need to manually correct that.

1. Open the table in example/POS/is_relative_quantification/forced_targeted_peak_table_temple.xlsx, like the below figure shows:

1. Then open the peak shape plot for each internal standard in example/POS/is_relative_quantification/peak_shape. Then if you find that the peak integration is not correct, you can add the right begin and end retention time for peaks like below video shows:

library(vembedr)
embed_url("https://www.youtube.com/watch?v=OiUdqkw5KZc&feature=youtu.be") %>% 
  use_align("center") %>% 
  use_bs_responsive()

After check all the internal standards which are not normal, please rename it as forced_targeted_peak_table_temple_manual.xlsx and rerun get_relative_quantification(), and note that set forced_targeted_peak_table_name as forced_targeted_peak_table_temple_manual.xlsx:

get_relative_quantification(
  path = "example/POS",
  forced_targeted_peak_table_name = "forced_targeted_peak_table_temple_manual.xlsx",
  output_path_name = "is_relative_quantification",
  targeted_table_name = "IS_info_new.xlsx",
  sample_info = sample_info_pos,
  targeted_table_type = "is",
  polarity = "positive",
  chol_rt = chol_rt,
  output_integrate = TRUE,
  output_eic = TRUE,
  ppm = 40,
  rt.tolerance = 180,
  threads = 5,
  rerun = FALSE
)

Then open the internal standard 18_1_d7_ Lyso PC in positive mode (example/POS/is_relative_quantification/peak_shape/18_1_d7_ Lyso PC.html), the peak detection is good for now:

htmltools::tags$iframe(
  title = "Internal standard",
  src = "../man/figures/figure33.html",
  width = "100%",
  height = "600",
  scrolling = "no",
  seamless = "seamless",
  frameBorder = "0"
)

For the negative mode, we can also manually check internal standards like this and rerun to get more accurate peak integration and relative quantification data.

Next step

Next we need to get the absolute quantification data of all lipids in all sample.

Please click here:

Step3: get absolute quantification data

jaspershen/lipidflow documentation built on March 7, 2021, 2:42 p.m.