OSCR

Metabolomic signatures of SSRI exposure during neural differentiation and correlation of lysophosphatidylcholines with early symptoms of neurodevelopmental disorders.

Code ↔ Paper

The paper beside its authors' code: matches between them have not been computed for this paper yet.

Paper

Loaded from Europe PMC by your browser, not stored by OSCR: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The authors' code

R Markdown · 635 lines · 24 KB · CC0-1.0

  1. ---
  2. title: "SSRI – Differential Metabolomics Analysis"
  3. author: "Abishek Arora"
  4. date: "10/06/2025"
  5. ---
  7. # Differential Metabolomics Analysis
  9. Mass spectrometry was performed on samples for cell lines exposed to fluoxetine hydrochloride (FH), citalopram hydrobromide (CH), sertraline hydrochloride (SH) and paroxetine hydrochloride (PH) for 5 and 28 days. The cell lines that were a part of this study were CTRL9II, ASD12BI, ASD17AII and AF22. Metabolite concentrations are reported in micromolar (µM). Differential metabolomics analysis was performed using a mixed linear model as stated in the sections that follow. For more information about the methods applied, please refer to the paper.
  11. ## Fluoxetine hydrochloride (FH)
  13. ### Day 5
  15. Comparisons were made based on treatment groups of fluoxetine hydrochloride exposure, namely "Control" and "FH" at Day 5, using a mixed linear model followed by correction for multiple comparisons using the FDR method
  17. ```{r}
  18. # Load package libraries and import the dataset to be analysed.
  19. rm(list = ls())
  20. library("tidyverse")
  21. library("nlme")
  22. FH_metabolome <- read.csv("add file path to metabo_conc.csv here")
  23. FH_metabolome <- FH_metabolome %>% filter(diff_time == "Day 5", exp == c("Control", "FH"))
  24. FH_metabolome$exp <- factor(FH_metabolome$exp, levels = c("Control", "FH"))
  25. list_metab <- split(FH_metabolome, FH_metabolome$metabo)
  27. results_df <- read.csv("add file path to metabo_list.csv here", row.names=1)
  29. # Load function for preparing dataframe of the model outputs:
  30. merge.all <- function(x, ..., by = "row.names") {
  31. L <- list(...)
  32. for (i in seq_along(L)) {
  33. x <- merge(x, L[[i]], by = by)
  34. rownames(x) <- x$Row.names
  35. x$Row.names <- NULL
  36. }
  37. return(x)
  38. }
  39. ```
  41. ```{r}
  42. # Define the linear model and loop through the summary results.
  43. results <- vector(mode = "list", length = 189)
  44. for (i in 1:length(list_metab)) {
  45. tryCatch({
  46. model <- lme(metabo_conc ~ exp, random = ~ 1 | cell_line, data = list_metab[[i]], na.action = na.omit)
  47. results[[i]] <- summary(model)
  48. }, error = function(e){cat("ERROR :",conditionMessage(e), "\n")})
  49. }
  51. # Aggregate the p values from the summary results for the linear model.
  52. pvalues <- vector(mode = "list", length = 189)
  53. for (i in 1:length(results)) {
  54. df <- as.data.frame(results[[i]][["tTable"]])
  55. pvalues[[i]] <- df[2,5]
  56. }
  58. names(pvalues) <- names(list_metab)
  59. summary_res <- unlist(pvalues)
  60. summary_res <- as.data.frame(summary_res, add.rownames = True)
  62. summary_results <- merge(x = results_df, y = summary_res, by = 'row.names', all.x = TRUE)
  63. summary_results <- summary_results %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  65. # Aggregate the estimates from the summary results for the linear model.
  66. ests <- vector(mode = "list", length = 189)
  67. for (i in 1:length(results)) {
  68. df <- as.data.frame(results[[i]][["tTable"]])
  69. ests[[i]] <- df[2,1]
  70. }
  72. names(ests) <- names(list_metab)
  73. est_res <- unlist(ests)
  74. est_res <- as.data.frame(est_res, add.rownames = True)
  76. est_res <- merge(x = results_df, y = est_res, by = 'row.names', all.x = TRUE)
  77. est_res <- est_res %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  78. est_res[,1] <- NULL
  80. # Perform correction for multiple comparisons using the FDR method.
  81. p.adj <- p.adjust(summary_results$summary_res, method = "fdr", n = length(summary_results$summary_res))
  82. names(p.adj) <- names(list_metab)
  83. p.adj.df <- as.data.frame(p.adj)
  85. # Combine all dataframes for final result.
  86. final_res <- merge.all(summary_results, est_res, p.adj.df)
  87. colnames(final_res) <- c("Category", "p value", "Estimate", "Adj. p value")
  88. write.csv(final_res,"add file path for desired folder to save results here", row.names = TRUE)
  89. ```
  91. ### Day 28
  93. Comparisons were made based on treatment groups of fluoxetine hydrochloride exposure, namely "Control" and "FH" at Day 28, using a mixed linear model followed by correction for multiple comparisons using the FDR method
  95. ```{r}
  96. # Load package libraries and import the dataset to be analysed.
  97. rm(list = ls())
  98. library("tidyverse")
  99. FH_metabolome <- read.csv("add file path to metabo_conc.csv here")
  100. FH_metabolome <- FH_metabolome %>% filter(diff_time == "Day 28", exp == c("Control", "FH"))
  101. FH_metabolome$exp <- factor(FH_metabolome$exp, levels = c("Control", "FH"))
  102. list_metab <- split(FH_metabolome, FH_metabolome$metabo)
  104. results_df <- read.csv("add file path to metabo_list.csv here", row.names=1)
  106. # Load function for preparing dataframe of the model outputs:
  107. merge.all <- function(x, ..., by = "row.names") {
  108. L <- list(...)
  109. for (i in seq_along(L)) {
  110. x <- merge(x, L[[i]], by = by)
  111. rownames(x) <- x$Row.names
  112. x$Row.names <- NULL
  113. }
  114. return(x)
  115. }
  116. ```
  118. ```{r}
  119. # Define the linear model and loop through the summary results.
  120. results <- vector(mode = "list", length = 189)
  121. for (i in 1:length(list_metab)) {
  122. tryCatch({
  123. model <- lme(metabo_conc ~ exp, random = ~ 1 | cell_line, data = list_metab[[i]], na.action = na.omit)
  124. results[[i]] <- summary(model)
  125. }, error = function(e){cat("ERROR :",conditionMessage(e), "\n")})
  126. }
  128. # Aggregate the p values from the summary results for the linear model.
  129. pvalues <- vector(mode = "list", length = 189)
  130. for (i in 1:length(results)) {
  131. df <- as.data.frame(results[[i]][["tTable"]])
  132. pvalues[[i]] <- df[2,5]
  133. }
  135. names(pvalues) <- names(list_metab)
  136. summary_res <- unlist(pvalues)
  137. summary_res <- as.data.frame(summary_res, add.rownames = True)
  139. summary_results <- merge(x = results_df, y = summary_res, by = 'row.names', all.x = TRUE)
  140. summary_results <- summary_results %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  142. # Aggregate the estimates from the summary results for the linear model.
  143. ests <- vector(mode = "list", length = 189)
  144. for (i in 1:length(results)) {
  145. df <- as.data.frame(results[[i]][["tTable"]])
  146. ests[[i]] <- df[2,1]
  147. }
  149. names(ests) <- names(list_metab)
  150. est_res <- unlist(ests)
  151. est_res <- as.data.frame(est_res, add.rownames = True)
  153. est_res <- merge(x = results_df, y = est_res, by = 'row.names', all.x = TRUE)
  154. est_res <- est_res %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  155. est_res[,1] <- NULL
  157. # Perform correction for multiple comparisons using the FDR method.
  158. p.adj <- p.adjust(summary_results$summary_res, method = "fdr", n = length(summary_results$summary_res))
  159. names(p.adj) <- names(list_metab)
  160. p.adj.df <- as.data.frame(p.adj)
  162. # Combine all dataframes for final result.
  163. final_res <- merge.all(summary_results, est_res, p.adj.df)
  164. colnames(final_res) <- c("Category", "p value", "Estimate", "Adj. p value")
  165. write.csv(final_res,"add file path for desired folder to save results here", row.names = TRUE)
  166. ```
  169. ## Citalopram
  171. ### Day 5
  173. Comparisons were made based on treatment groups of citalopram hydrobromide exposure, namely "Control" and "CH" at day 5, using a mixed linear model followed by correction for multiple comparisons using the FDR method
  175. ```{r}
  176. # Load package libraries and import the dataset to be analysed.
  177. rm(list = ls())
  178. library("tidyverse")
  179. CH_metabolome <- read.csv("add file path to metabo_conc.csv here")
  180. CH_metabolome <- CH_metabolome %>% filter(diff_time == "Day 5", exp == c("Control", "CH"))
  181. CH_metabolome$exp <- factor(CH_metabolome$exp, levels = c("Control", "CH"))
  182. list_metab <- split(CH_metabolome, CH_metabolome$metabo)
  184. results_df <- read.csv("add file path to metabo_list.csv here", row.names=1)
  186. # Load function for preparing dataframe of the model outputs:
  187. merge.all <- function(x, ..., by = "row.names") {
  188. L <- list(...)
  189. for (i in seq_along(L)) {
  190. x <- merge(x, L[[i]], by = by)
  191. rownames(x) <- x$Row.names
  192. x$Row.names <- NULL
  193. }
  194. return(x)
  195. }
  196. ```
  198. ```{r}
  199. # Define the linear model and loop through the summary results.
  200. results <- vector(mode = "list", length = 189)
  201. for (i in 1:length(list_metab)) {
  202. tryCatch({
  203. model <- lme(metabo_conc ~ exp, random = ~ 1 | cell_line, data = list_metab[[i]], na.action = na.omit)
  204. results[[i]] <- summary(model)
  205. }, error = function(e){cat("ERROR :",conditionMessage(e), "\n")})
  206. }
  208. # Aggregate the p values from the summary results for the linear model.
  209. pvalues <- vector(mode = "list", length = 189)
  210. for (i in 1:length(results)) {
  211. df <- as.data.frame(results[[i]][["tTable"]])
  212. pvalues[[i]] <- df[2,5]
  213. }
  215. names(pvalues) <- names(list_metab)
  216. summary_res <- unlist(pvalues)
  217. summary_res <- as.data.frame(summary_res, add.rownames = True)
  219. summary_results <- merge(x = results_df, y = summary_res, by = 'row.names', all.x = TRUE)
  220. summary_results <- summary_results %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  222. # Aggregate the estimates from the summary results for the linear model.
  223. ests <- vector(mode = "list", length = 189)
  224. for (i in 1:length(results)) {
  225. df <- as.data.frame(results[[i]][["tTable"]])
  226. ests[[i]] <- df[2,1]
  227. }
  229. names(ests) <- names(list_metab)
  230. est_res <- unlist(ests)
  231. est_res <- as.data.frame(est_res, add.rownames = True)
  233. est_res <- merge(x = results_df, y = est_res, by = 'row.names', all.x = TRUE)
  234. est_res <- est_res %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  235. est_res[,1] <- NULL
  237. # Perform correction for multiple comparisons using the FDR method.
  238. p.adj <- p.adjust(summary_results$summary_res, method = "fdr", n = length(summary_results$summary_res))
  239. names(p.adj) <- names(list_metab)
  240. p.adj.df <- as.data.frame(p.adj)
  242. # Combine all dataframes for final result.
  243. final_res <- merge.all(summary_results, est_res, p.adj.df)
  244. colnames(final_res) <- c("Category", "p value", "Estimate", "Adj. p value")
  245. write.csv(final_res,"add file path for desired folder to save results here", row.names = TRUE)
  246. ```
  248. ### Day 28
  250. Comparisons were made based on treatment groups of citalopram hydrobromide exposure, namely "Control" and "CH" at day 28, using a mixed linear model followed by correction for multiple comparisons using the FDR method
  252. ```{r}
  253. # Load package libraries and import the dataset to be analysed.
  254. rm(list = ls())
  255. library("tidyverse")
  256. CH_metabolome <- read.csv("add file path to metabo_conc.csv here")
  257. CH_metabolome <- CH_metabolome %>% filter(diff_time == "Day 28", exp == c("Control", "CH"))
  258. CH_metabolome$exp <- factor(CH_metabolome$exp, levels = c("Control", "CH"))
  259. list_metab <- split(CH_metabolome, CH_metabolome$metabo)
  261. results_df <- read.csv("add file path to metabo_list.csv here", row.names=1)
  263. # Load function for preparing dataframe of the model outputs:
  264. merge.all <- function(x, ..., by = "row.names") {
  265. L <- list(...)
  266. for (i in seq_along(L)) {
  267. x <- merge(x, L[[i]], by = by)
  268. rownames(x) <- x$Row.names
  269. x$Row.names <- NULL
  270. }
  271. return(x)
  272. }
  273. ```
  275. ```{r}
  276. # Define the linear model and loop through the summary results.
  277. results <- vector(mode = "list", length = 189)
  278. for (i in 1:length(list_metab)) {
  279. tryCatch({
  280. model <- lme(metabo_conc ~ exp, random = ~ 1 | cell_line, data = list_metab[[i]], na.action = na.omit)
  281. results[[i]] <- summary(model)
  282. }, error = function(e){cat("ERROR :",conditionMessage(e), "\n")})
  283. }
  285. # Aggregate the p values from the summary results for the linear model.
  286. pvalues <- vector(mode = "list", length = 189)
  287. for (i in 1:length(results)) {
  288. df <- as.data.frame(results[[i]][["tTable"]])
  289. pvalues[[i]] <- df[2,5]
  290. }
  292. names(pvalues) <- names(list_metab)
  293. summary_res <- unlist(pvalues)
  294. summary_res <- as.data.frame(summary_res, add.rownames = True)
  296. summary_results <- merge(x = results_df, y = summary_res, by = 'row.names', all.x = TRUE)
  297. summary_results <- summary_results %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  299. # Aggregate the estimates from the summary results for the linear model.
  300. ests <- vector(mode = "list", length = 189)
  301. for (i in 1:length(results)) {
  302. df <- as.data.frame(results[[i]][["tTable"]])
  303. ests[[i]] <- df[2,1]
  304. }
  306. names(ests) <- names(list_metab)
  307. est_res <- unlist(ests)
  308. est_res <- as.data.frame(est_res, add.rownames = True)
  310. est_res <- merge(x = results_df, y = est_res, by = 'row.names', all.x = TRUE)
  311. est_res <- est_res %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  312. est_res[,1] <- NULL
  314. # Perform correction for multiple comparisons using the FDR method.
  315. p.adj <- p.adjust(summary_results$summary_res, method = "fdr", n = length(summary_results$summary_res))
  316. names(p.adj) <- names(list_metab)
  317. p.adj.df <- as.data.frame(p.adj)
  319. # Combine all dataframes for final result.
  320. final_res <- merge.all(summary_results, est_res, p.adj.df)
  321. colnames(final_res) <- c("Category", "p value", "Estimate", "Adj. p value")
  322. write.csv(final_res,"add file path for desired folder to save results here", row.names = TRUE)
  323. ```
  325. ## Sertraline
  327. ### Day 5
  329. Comparisons were made based on treatment groups of sertraline hydrochloride exposure, namely "Control" and "SH" at day 5, using a mixed linear model followed by correction for multiple comparisons using the FDR method
  331. ```{r}
  332. # Load package libraries and import the dataset to be analysed.
  333. rm(list = ls())
  334. library("tidyverse")
  335. SH_metabolome <- read.csv("add file path to metabo_conc.csv here")
  336. SH_metabolome <- SH_metabolome %>% filter(diff_time == "Day 5", exp == c("Control", "SH"))
  337. SH_metabolome$exp <- factor(SH_metabolome$exp, levels = c("Control", "SH"))
  338. list_metab <- split(SH_metabolome, SH_metabolome$metabo)
  340. results_df <- read.csv("add file path to metabo_list.csv here", row.names=1)
  342. # Load function for preparing dataframe of the model outputs:
  343. merge.all <- function(x, ..., by = "row.names") {
  344. L <- list(...)
  345. for (i in seq_along(L)) {
  346. x <- merge(x, L[[i]], by = by)
  347. rownames(x) <- x$Row.names
  348. x$Row.names <- NULL
  349. }
  350. return(x)
  351. }
  352. ```
  354. ```{r}
  355. # Define the linear model and loop through the summary results.
  356. results <- vector(mode = "list", length = 189)
  357. for (i in 1:length(list_metab)) {
  358. tryCatch({
  359. model <- lme(metabo_conc ~ exp, random = ~ 1 | cell_line, data = list_metab[[i]], na.action = na.omit)
  360. results[[i]] <- summary(model)
  361. }, error = function(e){cat("ERROR :",conditionMessage(e), "\n")})
  362. }
  364. # Aggregate the p values from the summary results for the linear model.
  365. pvalues <- vector(mode = "list", length = 189)
  366. for (i in 1:length(results)) {
  367. df <- as.data.frame(results[[i]][["tTable"]])
  368. pvalues[[i]] <- df[2,5]
  369. }
  371. names(pvalues) <- names(list_metab)
  372. summary_res <- unlist(pvalues)
  373. summary_res <- as.data.frame(summary_res, add.rownames = True)
  375. summary_results <- merge(x = results_df, y = summary_res, by = 'row.names', all.x = TRUE)
  376. summary_results <- summary_results %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  378. # Aggregate the estimates from the summary results for the linear model.
  379. ests <- vector(mode = "list", length = 189)
  380. for (i in 1:length(results)) {
  381. df <- as.data.frame(results[[i]][["tTable"]])
  382. ests[[i]] <- df[2,1]
  383. }
  385. names(ests) <- names(list_metab)
  386. est_res <- unlist(ests)
  387. est_res <- as.data.frame(est_res, add.rownames = True)
  389. est_res <- merge(x = results_df, y = est_res, by = 'row.names', all.x = TRUE)
  390. est_res <- est_res %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  391. est_res[,1] <- NULL
  393. # Perform correction for multiple comparisons using the FDR method.
  394. p.adj <- p.adjust(summary_results$summary_res, method = "fdr", n = length(summary_results$summary_res))
  395. names(p.adj) <- names(list_metab)
  396. p.adj.df <- as.data.frame(p.adj)
  398. # Combine all dataframes for final result.
  399. final_res <- merge.all(summary_results, est_res, p.adj.df)
  400. colnames(final_res) <- c("Category", "p value", "Estimate", "Adj. p value")
  401. write.csv(final_res,"add file path for desired folder to save results here", row.names = TRUE)
  402. ```
  404. ### Day 28
  406. Comparisons were made based on treatment groups of sertraline hydrochloride exposure, namely "Control" and "SH" at day 28, using a mixed linear model followed by correction for multiple comparisons using the FDR method
  408. ```{r}
  409. # Load package libraries and import the dataset to be analysed.
  410. rm(list = ls())
  411. library("tidyverse")
  412. SH_metabolome <- read.csv("add file path to metabo_conc.csv here")
  413. SH_metabolome <- SH_metabolome %>% filter(diff_time == "Day 28", exp == c("Control", "SH"))
  414. SH_metabolome$exp <- factor(SH_metabolome$exp, levels = c("Control", "SH"))
  415. list_metab <- split(SH_metabolome, SH_metabolome$metabo)
  417. results_df <- read.csv("add file path to metabo_list.csv here", row.names=1)
  419. # Load function for preparing dataframe of the model outputs:
  420. merge.all <- function(x, ..., by = "row.names") {
  421. L <- list(...)
  422. for (i in seq_along(L)) {
  423. x <- merge(x, L[[i]], by = by)
  424. rownames(x) <- x$Row.names
  425. x$Row.names <- NULL
  426. }
  427. return(x)
  428. }
  429. ```
  431. ```{r}
  432. # Define the linear model and loop through the summary results.
  433. results <- vector(mode = "list", length = 189)
  434. for (i in 1:length(list_metab)) {
  435. tryCatch({
  436. model <- lme(metabo_conc ~ exp, random = ~ 1 | cell_line, data = list_metab[[i]], na.action = na.omit)
  437. results[[i]] <- summary(model)
  438. }, error = function(e){cat("ERROR :",conditionMessage(e), "\n")})
  439. }
  441. # Aggregate the p values from the summary results for the linear model.
  442. pvalues <- vector(mode = "list", length = 189)
  443. for (i in 1:length(results)) {
  444. df <- as.data.frame(results[[i]][["tTable"]])
  445. pvalues[[i]] <- df[2,5]
  446. }
  448. names(pvalues) <- names(list_metab)
  449. summary_res <- unlist(pvalues)
  450. summary_res <- as.data.frame(summary_res, add.rownames = True)
  452. summary_results <- merge(x = results_df, y = summary_res, by = 'row.names', all.x = TRUE)
  453. summary_results <- summary_results %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  455. # Aggregate the estimates from the summary results for the linear model.
  456. ests <- vector(mode = "list", length = 189)
  457. for (i in 1:length(results)) {
  458. df <- as.data.frame(results[[i]][["tTable"]])
  459. ests[[i]] <- df[2,1]
  460. }
  462. names(ests) <- names(list_metab)
  463. est_res <- unlist(ests)
  464. est_res <- as.data.frame(est_res, add.rownames = True)
  466. est_res <- merge(x = results_df, y = est_res, by = 'row.names', all.x = TRUE)
  467. est_res <- est_res %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  468. est_res[,1] <- NULL
  470. # Perform correction for multiple comparisons using the FDR method.
  471. p.adj <- p.adjust(summary_results$summary_res, method = "fdr", n = length(summary_results$summary_res))
  472. names(p.adj) <- names(list_metab)
  473. p.adj.df <- as.data.frame(p.adj)
  475. # Combine all dataframes for final result.
  476. final_res <- merge.all(summary_results, est_res, p.adj.df)
  477. colnames(final_res) <- c("Category", "p value", "Estimate", "Adj. p value")
  478. write.csv(final_res,"add file path for desired folder to save results here", row.names = TRUE)
  479. ```
  481. ## Paroxetine
  483. ### Day 5
  485. Comparisons were made based on treatment groups of paroxetine hydrochloride exposure, namely "Control" and "PH" at day 5, using a mixed linear model followed by correction for multiple comparisons using the FDR method
  487. ```{r}
  488. # Load package libraries and import the dataset to be analysed.
  489. rm(list = ls())
  490. library("tidyverse")
  491. PH_metabolome <- read.csv("add file path to metabo_conc.csv here")
  492. PH_metabolome <- PH_metabolome %>% filter(diff_time == "Day 5", exp == c("Control", "PH"))
  493. PH_metabolome$exp <- factor(PH_metabolome$exp, levels = c("Control", "PH"))
  494. list_metab <- split(PH_metabolome, PH_metabolome$metabo)
  496. results_df <- read.csv("add file path to metabo_list.csv here", row.names=1)
  498. # Load function for preparing dataframe of the model outputs:
  499. merge.all <- function(x, ..., by = "row.names") {
  500. L <- list(...)
  501. for (i in seq_along(L)) {
  502. x <- merge(x, L[[i]], by = by)
  503. rownames(x) <- x$Row.names
  504. x$Row.names <- NULL
  505. }
  506. return(x)
  507. }
  508. ```
  510. ```{r}
  511. # Define the linear model and loop through the summary results.
  512. results <- vector(mode = "list", length = 189)
  513. for (i in 1:length(list_metab)) {
  514. tryCatch({
  515. model <- lme(metabo_conc ~ exp, random = ~ 1 | cell_line, data = list_metab[[i]], na.action = na.omit)
  516. results[[i]] <- summary(model)
  517. }, error = function(e){cat("ERROR :",conditionMessage(e), "\n")})
  518. }
  520. # Aggregate the p values from the summary results for the linear model.
  521. pvalues <- vector(mode = "list", length = 189)
  522. for (i in 1:length(results)) {
  523. df <- as.data.frame(results[[i]][["tTable"]])
  524. pvalues[[i]] <- df[2,5]
  525. }
  527. names(pvalues) <- names(list_metab)
  528. summary_res <- unlist(pvalues)
  529. summary_res <- as.data.frame(summary_res, add.rownames = True)
  531. summary_results <- merge(x = results_df, y = summary_res, by = 'row.names', all.x = TRUE)
  532. summary_results <- summary_results %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  534. # Aggregate the estimates from the summary results for the linear model.
  535. ests <- vector(mode = "list", length = 189)
  536. for (i in 1:length(results)) {
  537. df <- as.data.frame(results[[i]][["tTable"]])
  538. ests[[i]] <- df[2,1]
  539. }
  541. names(ests) <- names(list_metab)
  542. est_res <- unlist(ests)
  543. est_res <- as.data.frame(est_res, add.rownames = True)
  545. est_res <- merge(x = results_df, y = est_res, by = 'row.names', all.x = TRUE)
  546. est_res <- est_res %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  547. est_res[,1] <- NULL
  549. # Perform correction for multiple comparisons using the FDR method.
  550. p.adj <- p.adjust(summary_results$summary_res, method = "fdr", n = length(summary_results$summary_res))
  551. names(p.adj) <- names(list_metab)
  552. p.adj.df <- as.data.frame(p.adj)
  554. # Combine all dataframes for final result.
  555. final_res <- merge.all(summary_results, est_res, p.adj.df)
  556. colnames(final_res) <- c("Category", "p value", "Estimate", "Adj. p value")
  557. write.csv(final_res,"add file path for desired folder to save results here", row.names = TRUE)
  558. ```
  560. ### Day 28
  562. Comparisons were made based on treatment groups of paroxetine hydrochloride exposure, namely "Control" and "PH" at day 28, using a mixed linear model followed by correction for multiple comparisons using the FDR method
  564. ```{r}
  565. # Load package libraries and import the dataset to be analysed.
  566. rm(list = ls())
  567. library("tidyverse")
  568. PH_metabolome <- read.csv("add file path to metabo_conc.csv here")
  569. PH_metabolome <- PH_metabolome %>% filter(diff_time == "Day 28", exp == c("Control", "PH"))
  570. PH_metabolome$exp <- factor(PH_metabolome$exp, levels = c("Control", "PH"))
  571. list_metab <- split(PH_metabolome, PH_metabolome$metabo)
  573. results_df <- read.csv("add file path to metabo_list.csv here", row.names=1)
  575. # Load function for preparing dataframe of the model outputs:
  576. merge.all <- function(x, ..., by = "row.names") {
  577. L <- list(...)
  578. for (i in seq_along(L)) {
  579. x <- merge(x, L[[i]], by = by)
  580. rownames(x) <- x$Row.names
  581. x$Row.names <- NULL
  582. }
  583. return(x)
  584. }
  585. ```
  587. ```{r}
  588. # Define the linear model and loop through the summary results.
  589. results <- vector(mode = "list", length = 189)
  590. for (i in 1:length(list_metab)) {
  591. tryCatch({
  592. model <- lme(metabo_conc ~ exp, random = ~ 1 | cell_line, data = list_metab[[i]], na.action = na.omit)
  593. results[[i]] <- summary(model)
  594. }, error = function(e){cat("ERROR :",conditionMessage(e), "\n")})
  595. }
  597. # Aggregate the p values from the summary results for the linear model.
  598. pvalues <- vector(mode = "list", length = 189)
  599. for (i in 1:length(results)) {
  600. df <- as.data.frame(results[[i]][["tTable"]])
  601. pvalues[[i]] <- df[2,5]
  602. }
  604. names(pvalues) <- names(list_metab)
  605. summary_res <- unlist(pvalues)
  606. summary_res <- as.data.frame(summary_res, add.rownames = True)
  608. summary_results <- merge(x = results_df, y = summary_res, by = 'row.names', all.x = TRUE)
  609. summary_results <- summary_results %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  611. # Aggregate the estimates from the summary results for the linear model.
  612. ests <- vector(mode = "list", length = 189)
  613. for (i in 1:length(results)) {
  614. df <- as.data.frame(results[[i]][["tTable"]])
  615. ests[[i]] <- df[2,1]
  616. }
  618. names(ests) <- names(list_metab)
  619. est_res <- unlist(ests)
  620. est_res <- as.data.frame(est_res, add.rownames = True)
  622. est_res <- merge(x = results_df, y = est_res, by = 'row.names', all.x = TRUE)
  623. est_res <- est_res %>% remove_rownames %>% column_to_rownames(var = "Row.names")
  624. est_res[,1] <- NULL
  626. # Perform correction for multiple comparisons using the FDR method.
  627. p.adj <- p.adjust(summary_results$summary_res, method = "fdr", n = length(summary_results$summary_res))
  628. names(p.adj) <- names(list_metab)
  629. p.adj.df <- as.data.frame(p.adj)
  631. # Combine all dataframes for final result.
  632. final_res <- merge.all(summary_results, est_res, p.adj.df)
  633. colnames(final_res) <- c("Category", "p value", "Estimate", "Adj. p value")
  634. write.csv(final_res,"add file path for desired folder to save results here", row.names = TRUE)
  635. ```

SSRI_Differential_Metabolomics.Rmd at commit cf3022b, under CC0-1.0 · at the source

Overview

Authors: Abishek Arora1,2,3, Kristina Vacy4,5, Cátia Marques6, Mihai-Ovidiu Degeratu1,2, Francesca Mastropasqua1,2, Jenny Humphrey1,2, Xuan Ye1,2, Marika Oksanen1,2, Peter Vuillermin4,7, Anne-Louise Ponsonby4,7, Ingela Lanekoff6,8, Kristiina Tammimies1,2, the Barwon Infant Study Investigator Group
ORCID iDs: Abishek Arora, Kristina Vacy, Mihai-Ovidiu Degeratu, Kristiina Tammimies
  1. Center of Neurodevelopmental Disorders (KIND), Centre for Psychiatry Research, Department of Women's and Children's Health, Karolinska Institutet, and Child and Adolescent Psychiatry, Stockholm Health Care Services, Region Stockholm, Stockholm, Sweden
  2. Astrid Lindgren Children’s Hospital, Karolinska University Hospital, Region Stockholm, Stockholm, Sweden
  3. Science for Life Laboratory, Karolinska Institutet, Stockholm, Sweden
  4. Florey Institute of Neuroscience and Mental Health, University of Melbourne, Victoria, Australia
  5. Melbourne School of Population and Global Health, University of Melbourne, Victoria, Australia
  6. Department of Chemistry for Life Sciences, Uppsala University, Uppsala, Sweden
  7. Murdoch Children’s Research Institute, Royal Children's Hospital, and Department of Paediatrics, University of Melbourne, Victoria, Australia
  8. Center of Excellence for the Chemical Mechanisms of Life, Uppsala University, Uppsala, Sweden
Institutions: Karolinska University Hospital (Sweden); Science for Life Laboratory (Sweden); Region Stockholm (Sweden); Karolinska Institutet (Sweden); Stockholm Health Care Services (Sweden); The University of Melbourne (Australia); Florey Institute of Neuroscience and Mental Health (Australia); Uppsala University (Sweden); Royal Children's Hospital (Australia); Murdoch Children's Research Institute (Australia)
Journal: EBioMedicine, volume 128, article 106291
Dates: received 30 August 2025; accepted 24 April 2026; published online 18 May 2026; in print June 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1016/j.ebiom.2026.106291 · PMID 42150307 · PMCID PMC13213235 · OpenAlex W7161595015
Open access: gold, a free copy (OpenAlex)
Status: code verified
Categories: genetics / omics (modality), human (organism), other condition (population)
Methods: Statistics, Smoothing, state filtering, decompositions, Machine learning
Keywords: Selective serotonin reuptake inhibitors, Induced pluripotent stem cells, Clinical cohort, Neurodevelopmental disorders, Metabolomics, Cellular assays
MeSH: Cell Differentiation*, Lysophosphatidylcholines*, Metabolome*, Metabolomics*, Neurodevelopmental Disorders*, Neurons*, Selective Serotonin Reuptake Inhibitors*, Biomarkers, Cell Line, Female, Humans, Induced Pluripotent Stem Cells, Mitochondria, Pregnancy, Reactive Oxygen Species (* major topic)
Topic: Tryptophan and brain disorders (Biological Psychiatry, Neuroscience), according to OpenAlex
Funding: European Commission; Swedish Research Council; HKH Crown Princess Lovisa Child Health Care Association; the Swedish Foundation for International Cooperation in Research and Higher Education; Karolinska Institutet; Swedish Foundation for Strategic Research; Brain Foundation; European Research Council
Citations: not cited yet (Europe PMC); 95 references in the paper

Abstract

Background: Selective serotonin reuptake inhibitors (SSRIs) are often prescribed during pregnancy. Epidemiological studies examining in-utero SSRI exposure and neurodevelopmental outcomes such as autism and ADHD have produced mixed results, in part due to challenges in accounting for underlying maternal mental health conditions and other confounding factors. The molecular pathways through which SSRIs may affect early neurodevelopment remain poorly understood.

Methods: We exposed neuroepithelial stem cells derived from four human induced pluripotent stem cell (iPSC) lines to fluoxetine, citalopram, sertraline, and paroxetine. We then assessed cellular viability, reactive oxygen species (ROS) levels, mitochondrial function using adenosine triphosphate (ATP) assays, and performed high-throughput metabolomics at two timepoints: proliferation and neural differentiation stages. The key metabolic findings were validated in the in-vitro model and in a complementary population-based cohort, the Barwon Infant Study, consisting of 1074 mother-child pairs with analysed cord-blood metabolomes.

Findings: Sertraline and paroxetine significantly decreased ROS and ATP levels in-vitro, indicating mitochondrial alteration. Metabolomic profiling revealed consistent elevation of three lysophosphatidylcholines (LPCs 16:0, 18:0, 18:1) in cells exposed to all SSRIs except citalopram. We further observed elevated LPC levels in the cord blood of infants prenatally exposed to SSRIs compared to those who were not (LPC 16:0 sn1: β = 0.173, 95% CI = 0.011–0.334, p = 0.037; LPC 16:0 sn2: β = 0.175, 95% CI = 0.007–0.343, p = 0.042 and LPC 18:0 sn2 β = 0.174, 95% CI = 0.018–0.330, p = 0.029), with a dose-dependent correlation to autism (LPC 16:0 sn2: overall p = 0.01) and ADHD-related symptoms (LPC 16:0 sn1: overall p = 0.007; LPC 16:0 sn2: overall p = 0.002) at age two.

Interpretation: These findings provide insights into SSRI-induced molecular changes in human iPSC derived neural cell cultures and highlight candidate metabolites that were validated in a clinical cohort. This may warrant their further investigation as indicators of SSRI exposure and emphasise the need for exploring prenatal SSRI exposure effects and neurodevelopmental outcomes in a wider context of other more well-established liability factors, including genetic background.

Funding: Vetenskapsrådet, Swedish Foundation for Strategic Research, Hjärnfonden, H.K.H. Kronprinsessan Lovisas förening för barnasjukvård, StratNeuro, Swedish Foundation for International Cooperation in Research and Higher Education, Karolinska Institutet and the European Research Council.

Reproduced under the paper's license (CC BY), from the paper cited above.

Repository

Its files are read in the Code ↔ Paper reader above.

Tammimies-Lab/SSRI-Metabolomics

License: CC0-1.0
State: the link answers, verified on 28 September 2026
Evidence: files inventoried
Commit: cf3022bd4a1c66348e9bac9417f16704d850ed92, 14 July 2026
Languages: R (1)
Size: 5 files, 1 script
Software Heritage: not archived
Found in: “Data sharing statement”
Holds: README, license file, 1 notebook
Not found: CITATION.cff, environment file, tests, continuous integration, documentation
Tools: nlme (1 file), tidyverse (1 file)
Availability: 1 check, the latest on 28 September 2026: the link answers
  • 28 September 2026: the link answers
3 files
At the source: github.com/Tammimies-Lab/SSRI-Metabolomics

The paper's code and data availability statement is in the Data section.

Tracing map

Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.

What the map holds:

  • 1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
  • 1 script, each with its path and the digest of its content;
  • no match between paragraphs and code yet;
  • neither the text of the paper nor the code itself.

Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.

Data

No dataset and no data link were found in the paper.

Data sharing statement

The data and code required for differential metabolomics analysis are available on GitHub (https://github.com/Tammimies-Lab/SSRI-Metabolomics). The raw data from the mass spectrometry-based metabolomics is available on MetaboLights (MTBLS12645) or available upon request from the corresponding author (). The data from the Barwon Infant Study are available under restricted access to protect participant confidentiality. Researchers may request access by contacting the BIS Steering Committee via Anne-Louise Ponsonby at the Florey Institute of Neuroscience and Mental Health (). Requests are considered based on scientific and ethical considerations, and approved applications require a collaborative research agreement. Deidentified data can be shared in either Stata or CSV file formats. Further details about the cohort, data access process, and project overview are available at: www.barwoninfantstudy.org.au.

Reproduced under the paper's license (CC BY), from the paper cited above.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 1, 28 September 2026: the first record

Recorded: type, language, journal, volume, pages, dates, 13 authors, 6 keywords, 15 MeSH terms, 8 funders, 94 references.

Cite

This paper

Arora, A., Vacy, K., Marques, C., Degeratu, M.-O., Mastropasqua, F., Humphrey, J., Ye, X., Oksanen, M., Vuillermin, P., Ponsonby, A.-L., Lanekoff, I., Tammimies, K., & the Barwon Infant Study Investigator Group. (2026). Metabolomic signatures of SSRI exposure during neural differentiation and correlation of lysophosphatidylcholines with early symptoms of neurodevelopmental disorders. EBioMedicine, 128, 106291. https://doi.org/10.1016/j.ebiom.2026.106291

BibTeX

@article{arora2026metabolomic,
author = {Arora, Abishek and Vacy, Kristina and Marques, Cátia and Degeratu, Mihai-Ovidiu and Mastropasqua, Francesca and Humphrey, Jenny and Ye, Xuan and Oksanen, Marika and Vuillermin, Peter and Ponsonby, Anne-Louise and Lanekoff, Ingela and Tammimies, Kristiina and {the Barwon Infant Study Investigator Group}},
title = {{Metabolomic signatures of SSRI exposure during neural differentiation and correlation of lysophosphatidylcholines with early symptoms of neurodevelopmental disorders}},
journal = {EBioMedicine},
year = {2026},
month = may,
volume = {128},
pages = {106291},
publisher = {Elsevier},
issn = {2352-3964},
doi = {10.1016/j.ebiom.2026.106291},
url = {https://doi.org/10.1016/j.ebiom.2026.106291},
pmid = {42150307},
pmcid = {PMC13213235}
}

RIS

TY - JOUR
AU - Arora, Abishek
AU - Vacy, Kristina
AU - Marques, Cátia
AU - Degeratu, Mihai-Ovidiu
AU - Mastropasqua, Francesca
AU - Humphrey, Jenny
AU - Ye, Xuan
AU - Oksanen, Marika
AU - Vuillermin, Peter
AU - Ponsonby, Anne-Louise
AU - Lanekoff, Ingela
AU - Tammimies, Kristiina
AU - the Barwon Infant Study Investigator Group
TI - Metabolomic signatures of SSRI exposure during neural differentiation and correlation of lysophosphatidylcholines with early symptoms of neurodevelopmental disorders
T2 - EBioMedicine
J2 - eBioMedicine
PY - 2026
DA - 2026/05/18
VL - 128
SP - 106291
SN - 2352-3964
PB - Elsevier
DO - 10.1016/j.ebiom.2026.106291
UR - https://doi.org/10.1016/j.ebiom.2026.106291
LA - en
ER -

CSL-JSON

{
"id": "10.1016/j.ebiom.2026.106291",
"type": "article-journal",
"title": "Metabolomic signatures of SSRI exposure during neural differentiation and correlation of lysophosphatidylcholines with early symptoms of neurodevelopmental disorders",
"container-title": "EBioMedicine",
"author": [
{
"family": "Arora",
"given": "Abishek"
},
{
"family": "Vacy",
"given": "Kristina"
},
{
"family": "Marques",
"given": "Cátia"
},
{
"family": "Degeratu",
"given": "Mihai-Ovidiu"
},
{
"family": "Mastropasqua",
"given": "Francesca"
},
{
"family": "Humphrey",
"given": "Jenny"
},
{
"family": "Ye",
"given": "Xuan"
},
{
"family": "Oksanen",
"given": "Marika"
},
{
"family": "Vuillermin",
"given": "Peter"
},
{
"family": "Ponsonby",
"given": "Anne-Louise"
},
{
"family": "Lanekoff",
"given": "Ingela"
},
{
"family": "Tammimies",
"given": "Kristiina"
},
{
"literal": "the Barwon Infant Study Investigator Group"
}
],
"container-title-short": "eBioMedicine",
"volume": "128",
"page": "106291",
"DOI": "10.1016/j.ebiom.2026.106291",
"PMID": "42150307",
"PMCID": "PMC13213235",
"ISSN": "2352-3964",
"publisher": "Elsevier",
"URL": "https://doi.org/10.1016/j.ebiom.2026.106291",
"language": "en",
"issued": {
"date-parts": [
[
2026,
5,
18
]
]
}
}

The tracing map gets a citation of its own once an author has validated it and it has a DOI.

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.1186/s12888-026-08178-8 [code]
Machine learning model to identify gut microbiome-derived metabolites as potential biomarkers of autism spectrum disorder: a pilot study.
Journal: BMC psychiatry
In common: genetics / omics, 3 references
[2] doi:10.64898/2026.08.13.26360304 [code]
Lifespan brain structural variation reveals shared organization across mental health conditions
Journal: medRxiv (preprint)
In common: author Kristiina Tammimies
[3] doi:10.1002/hbm.70605 [code]
BrainEnrich: Revealing Biological Insights for Imaging-Derived Phenotypes Through Transcriptomic Enrichment.
Journal: Human brain mapping
In common: nlme, tidyverse, genetics / omics
[4] doi:10.1038/s41562-026-02486-5 [code]
Genome-wide association studies of infant and toddler temperament in European and multi-ancestry populations.
Journal: Nature human behaviour
In common: nlme, tidyverse, genetics / omics
[5] doi:10.1038/s41467-026-71542-5 [code]
Astrocyte fatty acid metabolism as a driver of risk for major depressive disorder.
Journal: Nature communications
In common: nlme, tidyverse, genetics / omics
[6] doi:10.1038/s41380-026-03694-1 [code]
Targeting cortico-striatal-amygdalar networks via theta-band frontoparietal synchronization in opioid use disorder: a randomized tACS-fMRI Trial.
Journal: Molecular psychiatry
In common: nlme, tidyverse, other condition
[7] doi:10.1016/j.nicl.2026.104012 [code]
Structural-functional multilayer brain network properties and outcome of combined repetitive transcranial magnetic stimulation and psychotherapy for obsessive-compulsive disorder.
Journal: NeuroImage. Clinical
In common: nlme, tidyverse, other condition
[8] doi:10.1038/s41398-026-04122-2 [code]
Multi-session CBM-I for social anxiety: examining psychopathology, cognitive, neural, and psychophysiological effects in a randomized controlled trial.
Journal: Translational psychiatry
In common: nlme, tidyverse, other condition
[9] doi:10.1038/s41467-026-71415-x [code]
Regional BOLD variability reflects microstructural maturation and neuronal ensheathment in the preterm infant cortex.
Journal: Nature communications
In common: nlme, tidyverse, other condition
[10] doi:10.1002/npr2.70125
Glutathione-Related Metabolite Levels and Enzyme Activities in Depression: A Systematic Review and Meta-Analysis.
Journal: Neuropsychopharmacology reports
In common: 2 references

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion, reproductions, activity

Discussion: questions and error reports about this paper and its code, from signed-in readers and its authors. It opens with sign-in.

Reproductions: reports from readers who ran the authors' code: what they reproduced, with which environment, commit and data. It opens with sign-in.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.