OSCR

Mechanistically informed circulating biomarkers are associated with acquired epilepsy after neonatal brain injury.

Code ↔ Paper

4 matches between paragraphs of the paper and lines of its authors' code, computed by the harvester (lexical-v1). Click a colored paragraph or line to see its counterpart.

The 4 matches
  1. [1] § Materials and methods › Protein biomarker measurements ↔ neonatal_biomarkers_github.Rmd, lines 213–254 · score 0.93 · complement component, CX3CL1, growth factor, NCAM, VCAM, enolase
  2. [2] § Results › Protein biomarker analyses ↔ neonatal_biomarkers_github.Rmd, lines 355–391 · score 0.68 · anti inflammatory cytokines, pro inflammatory cytokines, growth factors, FDR adjusted, log10, peptides
  3. [3] § Materials and methods › Model evaluation with receiver operating characteristic (ROC) and precision-recall curves ↔ neonatal_biomarkers_github.Rmd, lines 256–287 · score 0.52 · EEG seizures, abnormal neurological, exam, discharge, neonatal, epilepsy
  4. [4] § Materials and methods › Study design and participants ↔ neonatal_biomarkers_github.Rmd, lines 256–287 · score 0.51 · therapeutic hypothermia, WIDEA, birth, gestation, age, seizures

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 · 443 lines · 15 KB · no license · 4 matches

  1. ---
  2. title: "HEAL EEG Biomarkers"
  3. author: ""
  4. date: "`r Sys.Date()`"
  5. output:
  6. html_document:
  7. toc: true
  8. toc_float: true
  9. number_sections: true
  10. theme: readable
  11. ---
  13. <style>
  14. img { max-width: none; }
  15. </style>
  17. ```{r setup, include=FALSE}
  18. knitr::opts_chunk$set(
  19. echo = TRUE,
  20. message = FALSE,
  21. warning = FALSE,
  22. fig.width = 10,
  23. fig.height = 6
  24. )
  26. rm(list = ls())
  27. gc()
  29. required_packages <- c(
  30. "knitr", "ggplot2", "tableone", "corrplot", "PerformanceAnalytics",
  31. "ggExtra", "EnvStats", "sandwich", "DescTools", "lmtest", "glmnet",
  32. "mice", "pracma", "dotwhisker", "redcapAPI", "data.table",
  33. "Hmisc", "dplyr", "writexl", "readxl", "REDCapR", "janitor",
  34. "ggrepel", "pathfindR", "org.Hs.eg.db"
  35. )
  37. missing_packages <- required_packages[!required_packages %in% rownames(installed.packages())]
  38. if (length(missing_packages) > 0) {
  39. message("Install missing packages before knitting: ", paste(missing_packages, collapse = ", "))
  40. }
  42. invisible(lapply(required_packages, require, character.only = TRUE))
  43. ```
  45. # Notes
  47. This document is saved as HTML because some figures are too wide to fit on a PDF page. If figures do not render well in the browser, download the HTML file and open it locally.
  49. This GitHub version has been sanitized for public sharing:
  51. - REDCap API tokens have been removed.
  52. - Local file paths have been replaced with configurable paths.
  53. - Private identifiers and institution-specific paths should be reviewed before posting.
  54. - Raw data are not included in this repository.
  56. # Configuration
  58. Create a local `.Renviron` file, not committed to GitHub, containing:
  60. ```bash
  61. REDCAP_URI="https://redcap.ucsf.edu/api/"
  62. REDCAP_TOKEN="your_redcap_token_here"
  63. ```
  65. Set paths below to match your local environment. For GitHub, place de-identified input files under `data/` and generated outputs under `outputs/`.
  67. ```{r config}
  68. paths <- list(
  69. cytokine_file = "data/Cytokines_for_NSR_RISE.xlsx",
  70. output_dir = "outputs"
  71. )
  73. if (!dir.exists(paths$output_dir)) {
  74. dir.create(paths$output_dir, recursive = TRUE)
  75. }
  77. redcap_uri <- Sys.getenv("REDCAP_URI")
  78. redcap_token <- Sys.getenv("REDCAP_TOKEN")
  79. ```
  81. # Data processing
  83. We identified three samples with unusual biomarker readings that may have been urine samples instead of blood. Those three samples, the second samples for CLA-101, CLA-201, and PIT-113, were removed from this analysis.
  85. For biomarker concentrations reported as thresholds, for example `< some value`, the threshold was used as the value.
  87. These analyses were updated with the re-run tau measurements.
  89. Throughout, biomarkers are ordered and grouped as follows:
  91. - Brain-specific proteins: GFAP, NSE, S100B, UCH-L1, Tau
  92. - Other growth factors: BDNF, EPO, NCAM1/CD56, NRG1β, VEGF
  93. - Pro-inflammatory cytokines, chemokines, and peptides: C5a, Fractalkine/CXCL1, ICAM1/CD54, IFN-γ, IL-1β, IL-6, IL-12p70, IL-17A, MCP-1/CCL2, MIP-1α/CCL3, MIP-1β/CCL4, VCAM1/CD106
  94. - Anti-inflammatory cytokines: IL-1RA, IL-8, IL-10, IL-13, IL-33
  96. ```{r import-data, eval=FALSE}
  97. # Option A: Export directly from REDCap.
  98. # Do not hard-code REDCap tokens in this file.
  100. if (redcap_uri == "" || redcap_token == "") {
  101. stop("Set REDCAP_URI and REDCAP_TOKEN in your local .Renviron file.")
  102. }
  104. rcon <- redcapConnection(url = redcap_uri, token = redcap_token)
  106. derived.data <- exportRecordsTyped(
  107. rcon,
  108. fields = NULL,
  109. forms = NULL,
  110. records = NULL,
  111. events = NULL,
  112. labels = TRUE,
  113. dates = TRUE,
  114. survey = TRUE,
  115. checkboxLabels = FALSE,
  116. factors = FALSE
  117. )
  119. cytokines <- readxl::read_excel(paths$cytokine_file, na = ".")
  120. derived.data <- merge(derived.data, cytokines, by = "study_id", all = TRUE)
  121. ```
  123. ```{r import-from-local, eval=FALSE}
  124. # Option B: Load a de-identified RDS exported from REDCap.
  125. # This is preferred for public GitHub workflows.
  127. derived.data <- readRDS("data/derived_data_deidentified.rds")
  128. cytokines <- readxl::read_excel(paths$cytokine_file, na = ".")
  129. derived.data <- merge(derived.data, cytokines, by = "study_id", all = TRUE)
  130. ```
  132. # Variable derivation
  134. ```{r variable-derivation, eval=FALSE}
  135. # Preserve one study ID with missing blood volume that was known to be valid.
  136. derived.data.extra <- subset(derived.data, study_id == "02308")
  137. derived.data <- subset(derived.data, vol_blood != "NA")
  138. derived.data <- bind_rows(derived.data, derived.data.extra)
  140. # Exclude participants excluded at 24 months.
  141. derived.data <- subset(derived.data, final_24_excluded != "Yes" | is.na(final_24_excluded))
  143. # Participant type.
  144. derived.data$patient_type <- as.factor(derived.data$patient_type)
  145. levels(derived.data$patient_type) <- list("Cases" = c("Acute"), "Controls" = c("Control"))
  147. # Demographics and clinical variables for Table 1.
  148. derived.data$`Maternal race` <- as.factor(derived.data$race)
  149. levels(derived.data$`Maternal race`) <- list(
  150. "White" = c("White"),
  151. "Black" = c("Black / African American"),
  152. "Asian" = c("Asian"),
  153. "Multiple/Other/Unknown" = c(
  154. "American Indian / Alaska Native",
  155. "Native Hawaiian / Other Pacific Islander",
  156. "Other",
  157. "Moret than one race",
  158. "Unknown / Not Reported",
  159. "Decline to Answer"
  160. )
  161. )
  163. derived.data$`Maternal ethnicity` <- as.factor(derived.data$ethnicity)
  164. levels(derived.data$`Maternal ethnicity`) <- list(
  165. "No/not reported" = c("NOT Hispanic or Latino", "Unknown / Not reported"),
  166. "Yes" = c("Hispanic or Latino")
  167. )
  169. derived.data$`Maternal education` <- as.factor(derived.data$maternal_edu)
  170. levels(derived.data$`Maternal education`) <- list(
  171. "At least some college or not reported" = c("College graduate", "Graduate study", "Unknown/Unavailable", "Decline to Answer"),
  172. "High school or less" = c("High school graduate", "Some education, high school not complete")
  173. )
  175. derived.data$`Insurance Type` <- as.factor(derived.data$insurance_type)
  176. levels(derived.data$`Insurance Type`) <- list(
  177. "Private" = c("Private"),
  178. "Public" = c("Public", "Unknown", "Decline to Answer", "None")
  179. )
  181. derived.data <- derived.data %>%
  182. mutate(`Maternal chorioamnionitis` = case_when(
  183. maternal_fever == "No Fever" ~ "No",
  184. maternal_fever %in% c(
  185. "Fever > 39C (102.2F) once",
  186. "Fever > 38C (100.4F) to 38.9C (102.0F) on two occasions"
  187. ) ~ "Yes",
  188. TRUE ~ NA_character_
  189. ))
  191. derived.data$`Cesarean section delivery` <- as.factor(derived.data$delivery_mode)
  192. levels(derived.data$`Cesarean section delivery`) <- list(
  193. "No" = c("Vaginal"),
  194. "Yes" = c("Operative vaginal (vacuum or forceps)", "Scheduled cesarean section", "Emergent cesarean section")
  195. )
  197. derived.data$`Birth location` <- as.factor(derived.data$location)
  198. levels(derived.data$`Birth location`) <- list(
  199. "Inborn" = c("Study center inpatient ward or ICU", "Study center L&D or well baby nursery"),
  200. "Other hospital/other" = c("Referral center L&D, well baby nursery, or ICU", "Home")
  201. )
  203. derived.data$Sex <- as.factor(derived.data$gender)
  204. levels(derived.data$Sex) <- list("Male" = c("Male"), "Female" = c("Female"))
  206. derived.data$`Birth weight, g` <- as.numeric(derived.data$birth_weight)
  207. derived.data$`Gestational age at birth, wk` <- as.numeric(derived.data$gest_age)
  208. derived.data$`5 min Apgar` <- as.numeric(derived.data$apgar_5min)
  209. derived.data$`10 min Apgar` <- as.numeric(derived.data$apgar_10min)
  210. derived.data$`Gestational age at neonatal blood draw, wk` <- as.numeric(derived.data$postmens_age_neobd)
  211. ```
  213. # Biomarker naming and ordering
  215. ```{r biomarker-names, eval=FALSE}
  216. # Rename raw biomarker columns to analysis-friendly names.
  217. # Keep this block synchronized with the input assay file.
  219. data.table::setnames(
  220. derived.data,
  221. old = c(
  222. 'hGFAP1','Tau1','Fractalkine1','IFNy1','IL-101','IL-12(p70)1','IL-131','IL-17A1','IL-1b1',
  223. 'IL-211','IL-231','IL-41','IL-51','IL-61','IL-81','MIP-1a1','MIP-1b1','TNFa1','BDNF1',
  224. 'CCL2/JE/MCP-11','Complement Component C5a1','Enolase 2/Neuron-specific Enolase1','Growth Hormone1',
  225. 'ICAM-1/CD541','IL-1ra/IL-1F31','IL-331','NCAM-1/CD561','Neuregulin-1 beta 1/NRG1 beta 11',
  226. 'S100B1','UCH-L1/PGP9.51','VCAM-1/CD1061','VEGF1'
  227. ),
  228. new = c(
  229. 'GFAP1','TAU1','CX3CL1/Fractalkine1','IFN-g1','IL-101','IL-12p701','IL-131','IL-171','IL-1b1',
  230. 'IL-211','IL-231','IL-41','IL-51','IL-61','IL-81','CCL3/MIP-1a1','CCL4/MIP-1b1','TNF-a1',
  231. 'BDNF1','CCL2/MCP11','C5a1','NSE1','GH1','CD54/ICAM-11','IL-1RA1','IL-331','CD56/NCAM11',
  232. 'NGB1','s100b1','UCHL11','CD106/VCAM-11','VEGF1'
  233. ),
  234. skip_absent = TRUE
  235. )
  237. biomarker_names_brain_specific_proteins <- c("GFAP1", "s100b1", "UCHL11", "TAU1")
  238. biomarker_names_other_growth_factors <- c("BDNF1", "CD56/NCAM11", "GH1", "NGB1", "NSE1", "VEGF1")
  239. biomarker_names_pro_inflammatory <- c(
  240. "C5a1", "CX3CL1/Fractalkine1", "CD54/ICAM-11", "IFN-g1", "IL-1b1", "IL-41", "IL-51",
  241. "IL-61", "IL-81", "IL-12p701", "IL-171", "IL-211", "IL-231", "CCL2/MCP11", "CCL3/MIP-1a1",
  242. "CCL4/MIP-1b1", "CD106/VCAM-11", "TNF-a1"
  243. )
  244. biomarker_names_anti_inflammatory <- c("IL-1RA1", "IL-101", "IL-131", "IL-331")
  246. biomarker_names <- c(
  247. biomarker_names_anti_inflammatory,
  248. biomarker_names_pro_inflammatory,
  249. biomarker_names_brain_specific_proteins,
  250. biomarker_names_other_growth_factors
  251. )
  253. biomarker_names_core <- unique(substr(biomarker_names, 1, nchar(biomarker_names) - 1))
  254. ```
  256. # Cohort tables
  258. ```{r cohort-tables, eval=FALSE}
  259. table_3_vars <- c(
  260. "Maternal race", "Maternal ethnicity", "Maternal education", "Insurance Type",
  261. "Maternal chorioamnionitis", "Cesarean section delivery", "Birth location", "Sex",
  262. "Birth weight, g", "Gestational age at birth, wk", "Gestational age at neonatal blood draw, wk",
  263. "5 min Apgar", "10 min Apgar", "Seizure Etiology", "Worst EEG Background",
  264. "Days of EEG Seizures", "Therapeutic Hypothermia", "Abnormal Neurologic Exam at Discharge",
  265. "CP", "WIDEA", "Services"
  266. )
  268. analysis.data <- derived.data %>%
  269. filter(!is.na(Epilepsy))
  271. table_3 <- CreateTableOne(
  272. data = analysis.data,
  273. vars = table_3_vars,
  274. strata = "Epilepsy",
  275. test = TRUE,
  276. addOverall = TRUE
  277. )
  279. table_3_print <- print(
  280. table_3,
  281. showAllLevels = FALSE,
  282. missing = FALSE,
  283. nonnormal = c("5 min Apgar", "10 min Apgar", "Gestational age at neonatal blood draw, wk")
  284. )
  286. knitr::kable(table_3_print, caption = "Cohort characteristics by epilepsy outcome")
  287. ```
  289. # Missing data
  291. ```{r missing-data, eval=FALSE}
  292. biomarker.data$number_biomarker <- rowSums(!is.na(biomarker.data[, biomarker_names]))
  294. table_missing_people <- data.frame(table(biomarker.data$number_biomarker))
  295. colnames(table_missing_people) <- c("Number of measures", "Number of participants")
  296. knitr::kable(table_missing_people, caption = "Number of biomarker measures available for participants")
  297. ```
  299. # Primary biomarker analysis
  301. ```{r primary-biomarker-analysis, eval=FALSE}
  302. manual_winsorize <- function(x, lower = 0.025, upper = 0.975) {
  303. qnt <- quantile(x, probs = c(lower, upper), na.rm = TRUE)
  304. ifelse(x < qnt[1], qnt[1], ifelse(x > qnt[2], qnt[2], x))
  305. }
  307. analysis.data$epilepsy <- ifelse(analysis.data$Epilepsy == 1, 1, 0)
  309. primary_df <- data.frame(
  310. biomarker = biomarker_names,
  311. median_no = NA_real_,
  312. iqr_no = NA_real_,
  313. median_yes = NA_real_,
  314. iqr_yes = NA_real_,
  315. rr = NA_real_,
  316. rr_ci_lower = NA_real_,
  317. rr_ci_upper = NA_real_,
  318. rr_p = NA_real_
  319. )
  321. primary_df$biomarker_name <- substr(primary_df$biomarker, 1, nchar(primary_df$biomarker) - 1)
  322. primary_df$sample <- substr(primary_df$biomarker, nchar(primary_df$biomarker), nchar(primary_df$biomarker))
  323. cutoff <- qnorm(1.95 / 2)
  325. for (i in seq_len(nrow(primary_df))) {
  326. biomarker.i <- primary_df$biomarker[i]
  328. primary_df$median_no[i] <- median(analysis.data[analysis.data$epilepsy == 0, biomarker.i], na.rm = TRUE)
  329. primary_df$iqr_no[i] <- IQR(analysis.data[analysis.data$epilepsy == 0, biomarker.i], na.rm = TRUE)
  330. primary_df$median_yes[i] <- median(analysis.data[analysis.data$epilepsy == 1, biomarker.i], na.rm = TRUE)
  331. primary_df$iqr_yes[i] <- IQR(analysis.data[analysis.data$epilepsy == 1, biomarker.i], na.rm = TRUE)
  333. analysis.data$predictor.i <- log2(analysis.data[, biomarker.i])
  334. model.i <- glm(epilepsy ~ predictor.i, family = poisson, data = analysis.data)
  336. cov.model.i <- sandwich::vcovHC(model.i, type = "HC0")
  337. std.err.i <- sqrt(diag(cov.model.i))["predictor.i"]
  339. primary_df$rr[i] <- exp(coef(model.i)["predictor.i"])
  340. primary_df$rr_ci_lower[i] <- exp(coef(model.i)["predictor.i"] - cutoff * std.err.i)
  341. primary_df$rr_ci_upper[i] <- exp(coef(model.i)["predictor.i"] + cutoff * std.err.i)
  342. primary_df$rr_p[i] <- lmtest::coeftest(model.i, vcov = cov.model.i)["predictor.i", "Pr(>|z|)"]
  343. }
  345. primary_df <- primary_df %>%
  346. mutate(
  347. p_fdr = p.adjust(rr_p, method = "fdr"),
  348. fold_change = median_yes / median_no,
  349. log2_fc = log2(fold_change)
  350. )
  352. knitr::kable(primary_df, caption = "Primary biomarker associations with epilepsy")
  353. ```
  355. # Manhattan plot
  357. ```{r manhattan-plot, eval=FALSE}
  358. primary_df <- primary_df %>%
  359. mutate(
  360. category = case_when(
  361. biomarker_name %in% substr(biomarker_names_brain_specific_proteins, 1, nchar(biomarker_names_brain_specific_proteins) - 1) ~ "Brain-specific proteins",
  362. biomarker_name %in% substr(biomarker_names_other_growth_factors, 1, nchar(biomarker_names_other_growth_factors) - 1) ~ "Other growth factors",
  363. biomarker_name %in% substr(biomarker_names_pro_inflammatory, 1, nchar(biomarker_names_pro_inflammatory) - 1) ~ "Pro-inflammatory cytokines, chemokines, and peptides",
  364. biomarker_name %in% substr(biomarker_names_anti_inflammatory, 1, nchar(biomarker_names_anti_inflammatory) - 1) ~ "Anti-inflammatory cytokines",
  365. TRUE ~ "Other"
  366. ),
  367. neg_log10_fdr = -log10(p_fdr)
  368. )
  370. manhattan_plot <- ggplot(primary_df, aes(x = factor(biomarker_name, levels = unique(biomarker_name)), y = neg_log10_fdr, color = category)) +
  371. geom_point(stroke = 1.5) +
  372. theme_bw() +
  373. theme(
  374. text = element_text(size = 16),
  375. panel.grid.minor = element_blank(),
  376. panel.grid.major.x = element_blank(),
  377. axis.title.y = element_text(size = 18.5, face = "bold"),
  378. axis.text.x = element_text(angle = 90, vjust = 0.5, hjust = 1)
  379. ) +
  380. labs(y = expression("FDR-adjusted p-value " ~ (-log[10] * " scale")), x = "")
  382. manhattan_plot
  384. ggsave(
  385. filename = file.path(paths$output_dir, "manhattan_plot_RISE.png"),
  386. plot = manhattan_plot,
  387. width = 12,
  388. height = 6,
  389. dpi = 600
  390. )
  391. ```
  393. # Volcano plot
  395. ```{r volcano-plot, eval=FALSE}
  396. volcano_df <- primary_df %>%
  397. mutate(
  398. gene_type = case_when(
  399. fold_change >= 1.24 & p_fdr <= 0.05 ~ "up",
  400. fold_change <= 0.75 & p_fdr <= 0.05 ~ "down",
  401. TRUE ~ "ns"
  402. ),
  403. delabel = if_else(gene_type != "ns", biomarker_name, NA_character_)
  404. )
  406. volcano_plot <- ggplot(volcano_df, aes(x = log2_fc, y = -log10(p_fdr), fill = gene_type, label = delabel)) +
  407. geom_point(shape = 21, color = "black") +
  408. geom_hline(yintercept = -log10(0.05), linetype = "dashed", linewidth = 0.4) +
  409. geom_vline(xintercept = c(log2(0.75), log2(1.25)), linetype = "dashed", linewidth = 0.2) +
  410. ggrepel::geom_text_repel(size = 3, max.overlaps = Inf) +
  411. theme_bw() +
  412. labs(x = expression(log[2]("fold change")), y = expression("FDR-adjusted p-value " ~ (-log[10] * " scale")))
  414. volcano_plot
  416. ggsave(
  417. filename = file.path(paths$output_dir, "volcano_RISE.png"),
  418. plot = volcano_plot,
  419. width = 6,
  420. height = 3,
  421. dpi = 600
  422. )
  423. ```
  425. # Protein pathway analysis
  427. ```{r protein-pathway-analysis, eval=FALSE}
  428. pathway_input <- primary_df %>%
  429. transmute(
  430. Gene_symbol = biomarker_name,
  431. log2FC = log2_fc,
  432. P.Value = rr_p
  433. )
  435. pathway_output <- pathfindR::run_pathfindR(pathway_input, pin_name_path = "KEGG")
  436. pathfindR::enrichment_chart(pathway_output, top_terms = 10)
  437. ```
  439. # Session information
  441. ```{r session-info}
  442. sessionInfo()
  443. ```

neonatal_biomarkers_github.Rmd at commit b2bd9fa, no license · at the source

Overview

Authors: Adam L Numis1,2, Renée A Shellhaas3, Janet S Soul4, Marisa A Gardner1,2, Courtney J Wusthoff5, Giulia M Benedetti6, Clara Di Germanio7,8, Theo K Bammler9,10, David J Erle8, Walter L Eckalbar8, Charles E McCulloch11, Patrick J Heagerty10, Thomas R Wood12, Yvonne W Wu1,2, Sandra E Juul12, Daniel H Lowenstein1, Hannah C Glass1,2,11, the Neonatal Seizure Registry Study Group, HEAL Consortium
ORCID iDs: Marisa A Gardner, Walter L Eckalbar, Hannah C Glass
  1. Department of Neurology and Weill Institute for Neurosciences, University of California San Francisco, San Francisco, CA 94143 USA
  2. Department of Pediatrics, University of California San Francisco, San Francisco, CA 94143 USA
  3. Department of Neurology, Washington University in St. Louis, St. Louis, MO 63110 USA
  4. Department of Neurology, Boston Children’s Hospital and Harvard Medical School, Boston, MA 02115 USA
  5. Department of Neurology, University of California Davis, Davis, CA 95616 USA
  6. Department of Pediatrics, Division of Pediatric Neurology, University of Michigan, Ann Arbor, MI 48108 USA
  7. Vitalant Research Institute, Core Immunology Laboratory, San Francisco, CA 94105 USA
  8. Department of Medicine, University of California, San Francisco, CA 94143 USA
  9. Department of Environmental and Occupational Health Sciences, University of Washington, Seattle, WA 98195 USA
  10. Department of Biostatistics, University of Washington, Seattle, WA 98195 USA
  11. Department of Epidemiology and Biostatistics, University of California San Francisco, San Francisco, CA 94143 USA
  12. Department of Pediatrics, University of Washington, Seattle, WA 98195 USA
Institutions: University of California, San Francisco (United States); Washington University in St. Louis (United States); Boston Children's Hospital (United States); Harvard University (United States); University of California, Davis (United States); University of Michigan (United States); Michigan Medicine (United States); Vitalant Research Institute (United States); University of Washington (United States)
Journal: Journal of neuroinflammation, volume 23, issue 1, article 215
Dates: received 20 March 2026; accepted 30 April 2026; published online 9 May 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1186/s12974-026-03853-9 · PMID 42106818 · PMCID PMC13312635 · OpenAlex W7160695465
Open access: gold, a free copy (OpenAlex)
Status: code verified
Categories: EEG (modality), human (organism), epilepsy (population)
Methods: Connectivity, Statistics, Smoothing, state filtering, decompositions, Machine learning
Keywords: Epileptogenesis, Inflammation, Cytokines, MicroRNA, Neonatal Seizures, EEG, Epilepsy
MeSH: Biomarkers*, Brain Injuries*, Epilepsy*, Cohort Studies, Female, Humans, Infant, Newborn, Male, MicroRNAs, Prospective Studies, Ubiquitin Thiolesterase (* major topic)
Topic: Neonatal and fetal brain pathology (Pediatrics, Perinatology and Child Health, Medicine), according to OpenAlex
Funding: National Institute of Neurological Disorders and Stroke (K23NS105918); NINDS NIH HHS (K23NS105918)
Citations: not cited yet (Europe PMC); 94 references in the paper

Abstract

Background: Acute provoked neonatal seizures are a major risk factor for acquired epilepsy, yet clinicians lack reliable tools to identify neonates at highest risk. Preclinical data implicate innate immune activation and neuronal injury as key drivers of epileptogenesis, suggesting blood-based biomarkers could provide mechanistic insight and prognostic utility. We sought to identify biomarkers of epileptogenesis in neonates with acute provoked seizures after brain injury using multicenter cohorts.

Methods: We conducted a prospective, multi-cohort analysis across two independent studies. NSR-RISE enrolled neonates with EEG-confirmed acute provoked seizures of diverse etiologies. The HEAL trial enrolled neonates with hypoxic-ischemic encephalopathy; analyses were limited to those with seizures. Plasma proteins were quantified 48—96 h after seizure onset. Associations with acquired epilepsy by 24-months were evaluated using log-link models with robust standard errors and false-discovery rate correction (FDR < 0.05). Significant proteins were added to models including established clinical predictors (≥ 3 days of EEG seizures and abnormal neurological examination at discharge). Exploratory pathway enrichment used KEGG databases. NSR-RISE participants also underwent plasma microRNA (miRNA) sequencing with integrative pathway analyses.

Results: Among 35 neonates in NSR-RISE, 7 (20%) developed epilepsy; among 40 neonates in HEAL, 6 (15%) developed epilepsy. Across both cohorts, neonates with epilepsy had higher concentrations of the pro-inflammatory cytokine IL-1β and the neuronal injury marker UCHL1 compared to those without epilepsy. Growth hormone (GH), measured only in NSR-RISE, was decreased in neonates with epilepsy. Incorporation of biomarkers improved prognostic accuracy for epilepsy beyond clinical features alone (Area under the precision-recall curve (AUPRC) 0.30 (95%CI, 0.28—0.32) versus 0.91 (95%CI, 0.89—0.92); p < 0.001). Pathway enrichment analyses implicated innate immune signaling, including TLR/IL1/NF-κB-related and MAPK-associated IL-17 signaling. miRNA profiling identified 11 species differentially expressed between neonates with and without epilepsy, including brain-enriched miRNAs. Network analysis identified a co-expression module enriched for let-7f-5p and miR-146a-5p targeting TLR/IL1/NF-κB, MAPK, and JAK/STAT pathways.

Conclusions: Across two cohorts, mechanistically informed biomarkers were associated with acquired epilepsy after neonatal seizures. IL-1β, UCHL1, and GH reflect inflammation, neuronal injury, and impaired trophic signaling, while circulating miRNAs provide complementary mechanistic insight. Findings support a translational biomarker panel and highlight inflammation as a biologically plausible therapeutic target.

Supplementary Information: The online version contains supplementary material available at 10.1186/s12974-026-03853-9.

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

Repository

Its files are read in the Code ↔ Paper reader above, with 4 matches between paragraphs and lines of code.

aln142/NSR-RISE

License: none: the authors keep all their rights
State: the link answers, verified on 28 September 2026
Evidence: files inventoried
Commit: b2bd9fa1cc9f9989a04524c7e6a0af8eadc6a20f, 28 April 2026
Languages: R (1)
Size: 2 files, 1 script
Software Heritage: not archived
Found in: “Data availability”
Holds: 1 notebook
Not found: README, license file, CITATION.cff, environment file, tests, continuous integration, documentation
Tools: data.table (1 file), ggplot2 (1 file), glmnet (1 file), tidyverse (1 file)
Availability: 1 check, the latest on 28 September 2026: the link answers
  • 28 September 2026: the link answers
1 file
At the source: github.com/aln142/NSR-RISE.git

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;
  • 4 matches between paragraphs of the paper and lines of the code (method lexical-v1);
  • 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 Availability Statement

Data are available from the corresponding author upon request, subject to approvals. De-identified NSR-RISE and HEAL data are shared per NIH and institutional policies. Analytic code is available on GitHub at https://github.com/aln142/NSR-RISE.

The minimal dataset necessary to interpret and replicate the findings is not publicly available due to protection of participant privacy and institutional data-sharing policies involving human subjects research. De-identified datasets generated or analyzed during the current study are available from the corresponding author upon reasonable request and subject to institutional approvals and data use agreements. Analytic code supporting the findings of this study is publicly available at: https://github.com/aln142/NSR-RISE.git

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, issue, pages, dates, 19 authors, 7 keywords, 11 MeSH terms, 2 funders, 91 references.

Cite

This paper

Numis, A. L., Shellhaas, R. A., Soul, J. S., Gardner, M. A., Wusthoff, C. J., Benedetti, G. M., Di Germanio, C., Bammler, T. K., Erle, D. J., Eckalbar, W. L., McCulloch, C. E., Heagerty, P. J., Wood, T. R., Wu, Y. W., Juul, S. E., Lowenstein, D. H., Glass, H. C., the Neonatal Seizure Registry Study Group, & HEAL Consortium. (2026). Mechanistically informed circulating biomarkers are associated with acquired epilepsy after neonatal brain injury. Journal of neuroinflammation, 23(1), 215. https://doi.org/10.1186/s12974-026-03853-9

BibTeX

@article{numis2026mechanistically,
author = {Numis, Adam L and Shellhaas, Renée A and Soul, Janet S and Gardner, Marisa A and Wusthoff, Courtney J and Benedetti, Giulia M and Di Germanio, Clara and Bammler, Theo K and Erle, David J and Eckalbar, Walter L and McCulloch, Charles E and Heagerty, Patrick J and Wood, Thomas R and Wu, Yvonne W and Juul, Sandra E and Lowenstein, Daniel H and Glass, Hannah C and {the Neonatal Seizure Registry Study Group} and {HEAL Consortium}},
title = {{Mechanistically informed circulating biomarkers are associated with acquired epilepsy after neonatal brain injury}},
journal = {Journal of neuroinflammation},
year = {2026},
month = may,
volume = {23},
number = {1},
pages = {215},
publisher = {BMC},
issn = {1742-2094},
doi = {10.1186/s12974-026-03853-9},
url = {https://doi.org/10.1186/s12974-026-03853-9},
pmid = {42106818},
pmcid = {PMC13312635}
}

RIS

TY - JOUR
AU - Numis, Adam L
AU - Shellhaas, Renée A
AU - Soul, Janet S
AU - Gardner, Marisa A
AU - Wusthoff, Courtney J
AU - Benedetti, Giulia M
AU - Di Germanio, Clara
AU - Bammler, Theo K
AU - Erle, David J
AU - Eckalbar, Walter L
AU - McCulloch, Charles E
AU - Heagerty, Patrick J
AU - Wood, Thomas R
AU - Wu, Yvonne W
AU - Juul, Sandra E
AU - Lowenstein, Daniel H
AU - Glass, Hannah C
AU - the Neonatal Seizure Registry Study Group
AU - HEAL Consortium
TI - Mechanistically informed circulating biomarkers are associated with acquired epilepsy after neonatal brain injury
T2 - Journal of neuroinflammation
J2 - J Neuroinflammation
PY - 2026
DA - 2026/05/09
VL - 23
IS - 1
SP - 215
SN - 1742-2094
PB - BMC
DO - 10.1186/s12974-026-03853-9
UR - https://doi.org/10.1186/s12974-026-03853-9
LA - en
ER -

CSL-JSON

{
"id": "10.1186/s12974-026-03853-9",
"type": "article-journal",
"title": "Mechanistically informed circulating biomarkers are associated with acquired epilepsy after neonatal brain injury",
"container-title": "Journal of neuroinflammation",
"author": [
{
"family": "Numis",
"given": "Adam L"
},
{
"family": "Shellhaas",
"given": "Renée A"
},
{
"family": "Soul",
"given": "Janet S"
},
{
"family": "Gardner",
"given": "Marisa A"
},
{
"family": "Wusthoff",
"given": "Courtney J"
},
{
"family": "Benedetti",
"given": "Giulia M"
},
{
"family": "Di Germanio",
"given": "Clara"
},
{
"family": "Bammler",
"given": "Theo K"
},
{
"family": "Erle",
"given": "David J"
},
{
"family": "Eckalbar",
"given": "Walter L"
},
{
"family": "McCulloch",
"given": "Charles E"
},
{
"family": "Heagerty",
"given": "Patrick J"
},
{
"family": "Wood",
"given": "Thomas R"
},
{
"family": "Wu",
"given": "Yvonne W"
},
{
"family": "Juul",
"given": "Sandra E"
},
{
"family": "Lowenstein",
"given": "Daniel H"
},
{
"family": "Glass",
"given": "Hannah C"
},
{
"literal": "the Neonatal Seizure Registry Study Group"
},
{
"literal": "HEAL Consortium"
}
],
"container-title-short": "J Neuroinflammation",
"volume": "23",
"issue": "1",
"page": "215",
"DOI": "10.1186/s12974-026-03853-9",
"PMID": "42106818",
"PMCID": "PMC13312635",
"ISSN": "1742-2094",
"publisher": "BMC",
"URL": "https://doi.org/10.1186/s12974-026-03853-9",
"language": "en",
"issued": {
"date-parts": [
[
2026,
5,
9
]
]
}
}

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:10.3390/ijms27104466 [code]
Uncovering the Key Circuit FOSL2/FOS/EGR3/EGR1, Contributing to the Hyperexcitability of Excitatory Neurons in the Epileptic Temporal Cortex and Hippocampus.
Journal: International journal of molecular sciences
In common: glmnet, data.table, ggplot2, 1 other tool, epilepsy
[2] doi:10.1038/s41586-026-10877-x [code]
Human brain organoids record the passage of time over multiple years.
Journal: Nature
In common: glmnet, data.table, ggplot2, 1 other tool, 1 reference
[3] doi:10.1038/s41598-026-51501-2 [code]
Expanding canonical cortical cell type markers in the era of single-cell transcriptomics.
Journal: Scientific reports
In common: glmnet, data.table, ggplot2, 1 other tool, 1 reference
[4] doi:10.1038/s41467-026-70374-7 [code]
scTWAS: a powerful statistical framework for single-cell transcriptome-wide association studies.
Journal: Nature communications
In common: glmnet, data.table, ggplot2, 1 other tool, 1 reference
[5] doi:10.1038/s42003-026-10957-8 [code]
Brain defence by the extracellular matrix protein Cochlin.
Journal: Communications biology
In common: glmnet, data.table, ggplot2, 1 other tool
[6] doi:10.1126/sciadv.aec9291 [code]
Computational mechanisms of perception in autism revealed using games inspired by rodent operant tasks.
Journal: Science advances
In common: glmnet, data.table, ggplot2, 1 other tool
[7] doi:10.1093/neuonc/noag128 [code]
Spatially-resolved single-cell imaging of melanoma brain metastases identifies localized immune patterns predictive of immune checkpoint blockade response.
Journal: Neuro-oncology
In common: glmnet, data.table, ggplot2, 1 other tool
[8] doi:10.1038/s41467-026-76762-3 [code]
PD-1 regulates CD4&lt;sup&gt;+&lt;/sup&gt; T cell-mediated CD8&lt;sup&gt;+&lt;/sup&gt; T cell responses in the brain to balance viral control and neuroinflammation.
Journal: Nature communications
In common: glmnet, data.table, ggplot2, 1 other tool
[9] doi:10.1126/sciadv.aeg3223 [code]
The extreme diversity of retinal amacrine cells has deep evolutionary roots.
Journal: Science advances
In common: glmnet, data.table, ggplot2, 1 other tool
[10] doi:10.3390/ijms27156925 [code]
XGBoost-SHAP Interpretable Modeling Identifies and Validates an Eight-Gene Biomarker for Hepatic Encephalopathy Risk Prediction in Cirrhosis.
Journal: International journal of molecular sciences
In common: glmnet, data.table, ggplot2, 1 other tool

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.