Publications

Identification Of Dysregulated Transcription Factor Activity In Temporal Lobe Epilepsy

Published in medRxiv (preprint), 2025

Approximately a third of patients with epilepsy continue to have uncontrolled seizures despite appropriate treatment with available antiseizure medications (i.e. drug-resistant epilepsy). There is emerging evidence that transcription factor (TF) activity is both dysregulated in epilepsy and modifiable by small-molecule drugs, providing an opportunity for treatment innovation. Methods for identifying dysregulated TFs and their target genes are still in their nascent stage and the reproducibility of findings remains unclear. We aimed to determine the concordance of findings, in terms of the TFs dysregulated, across different studies of drug-resistant epilepsy and to evaluate the performance of different methods for identifying dysregulated TFs. We used publicly available single-nucleotide RNA-seq data to construct discovery and validation datasets comprising individuals with drug-resistant temporal lobe epilepsy and healthy controls. We found good concordance (83%, 105/126) between the pySCENIC (Python implemented Single-cell Regulatory Network Inference and Clustering) and hdWGCNA (high-dimensional Weighted Gene Co-expression Network Analysis) methods. Cell-type specific concordance across the discovery and validation datasets was low (36% 137/377) and this could be attributed, in part, to differences in data quality. In contrast, we found strong concordance between TFs that met strict concordance criteria in the current study with those implicated in a tissue-level study in patients with drug-resistant epilepsies, with the overlap being higher for TLE-related modules relative to modules for other drug-resistant epilepsies [86% (32/37) vs. 21% (18/84), Fisher’s exact test: 95% 7.40 to 85.5, p < 0.0001]. Most TFs identified had been reported as being associated with epilepsy in the overall literature (91%, 53/58). Our findings strengthen the hypothesis that TFs are key to the pathophysiology of drug-resistant epilepsy and could represent novel drug targets. We recommend that multiple methods be applied to optimise discovery.

Recommended citation: Zeibich, Robert; O’Brien, Terence J.; Perucca, Piero; Kwan, Patrick; Anderson, Alison. Identification of dysregulated transcription factor activity in temporal lobe epilepsy. Preprint: medRxiv. 2025 Apr 24. View article online.
Download Paper

Development Of A Machine Learning Polyvariant Risk Prediction Model For Severe Cutaneous Adverse Drug Reactions To Carbamazepine And Other Aromatic Antiseizure Medications

Published in THE LANCET (preprint), 2024

Importance. Carbamazepine (CBZ) is a first-line treatment for epilepsy and neuropathic pain but carries risk for severe and sometimes life-threatening cutaneous reactions. Across Asian populations, the human leukocyte antigen HLA-B*15:02 in chromosome 6 is strongly associated with these adverse reactions, but the estimated positive predictive value (PPV) is only 0.07.

Objective. To develop a polyvariant risk prediction model with improved PPV.

Design, Settings and Participants. Case-control study. Cases had severe cutaneous reactions (Stevens-Johnson syndrome [SJS] or toxic epidermal necrolysis [TEN]) induced by carbamazepine (n=84) or other aromatic antiseizure medications (n=28). Cases were matched with drug-tolerant controls (n=80). All cases and controls were of Han Chinese descent, and 56% (107/192) carried HLA-B*15:02.

Exposures. Carbamazepine and other aromatic antiseizure medications.

Main Outcomes and Measures. All participants underwent whole genome sequencing. Putative genetic variants that modify the risk of developing carbamazepine-induced SJS/TEN were identified using a machine-learning workflow. Different model types were evaluated based on the weighted F1 (WTD F1) and area under the receiver operating characteristic curve (AUC) scores, and using genotypes within the known risk region (chromosome 6). The best-performing model was then used to screen for additional putative genetic markers. The bootstrapped PPV value obtained from the resulting polyvariant model was compared to that estimated from HLA-B*15:02 carrier status.

Results. We identified 465 common variants that collectively achieved a median AUC of 0.98 (95% confidence interval [CI]: 0.9799-0.9828) in predicing carbamazepine-induced SJS/TEN. Bootstrap analysis showed the model had a median PPV of 0.20 (95% CI: 0.19-0.22) and a negative predictive value of 1.0. In cases of SJS/TEN induced by aromatic medications other than CBZ, the model achieved AUC of 0.82 (95% CI: 0.82-0.83) and WTD F1 of 0.76 (95% CI: 0.76-0.77), indicating good generalisability.

Conclusion and Relevance. A machine learning approach is able to enhance the genetic prediction of severe cutaneous reactions induced by aromatic antiseizure medications. Improved precision of genetic risk prediction could stratify HLA-B*15:02 carriers to receive CBZ treatment with increased safety.

Recommended citation: Zeibich, Robert; Anderson, Alison; Chen, Zhibin; Shi, Yi-Wu; Ng, Ching-Ching; Baum, Larry; Cherny, Stacey; Sham, Pak-Chung; Lim, Kheng Seang; Liao, Wei-Ping; O’Brien, Terence J.; Perucca, Piero; and Kwan, Patrick. Development of a Machine Learning Polyvariant Risk Prediction Model for Severe Cutaneous Adverse Drug Reactions to Carbamazepine and Other Aromatic Antiseizure Medications. Preprint: THE LANCET. 2024 Nov 9. View article online
Download Paper

Applications For Deep Learning In Epilepsy Research

Published in International Journal of Molecular Sciences, 2023

Epilepsy is a group of brain disorders characterised by an enduring predisposition to generate unprovoked seizures. Fuelled by advances in sequencing technologies and computational approaches, more than 900 genes have now been implicated in epilepsy. The development and optimisation of tools and methods for analysing the vast quantity of genomic data is a rapidly evolving area of research. Deep learning (DL) is a subset of machine learning (ML) that brings opportunity for novel investigative strategies that can be harnessed to gain new insights into the genomic risk of people with epilepsy. DL is being harnessed to address limitations in accuracy of long-read sequencing technologies, which improve on short-read methods. Tools that predict the functional consequence of genetic variation can represent breaking ground in addressing critical knowledge gaps, while methods that integrate independent but complimentary data enhance the predictive power of genetic data. We provide an overview of these DL tools and discuss how they may be applied to the analysis of genetic data for epilepsy research.

Recommended citation: Zeibich R, Kwan P, J O'Brien T, Perucca P, Ge Z, Anderson A. Applications for Deep Learning in Epilepsy Genetic Research. Int J Mol Sci. 2023 Sep 27. View article online.
Download Paper

Optimising Whole-Genome Sequencing Diagnostics In Epilepsy

Published in undefined, 2025

Background. Whole-genome sequencing (WGS) is increasingly applied in clinical medicine, but best practices for variant detection, curation and interpretation have yet to be determined. We investigated how the choice of software methods and tools can impact WGS diagnostic yield.

Methods. WGS data from 70 patients enrolled in a randomised controlled trial had been processed using a commonly applied ‘gold standard’ workflow. Variants were curated and classified in accordance with the American College of Medical Genetics (ACMG) guidelines by a multidisciplinary team (MDT) comprising bioinformaticians, neurologists, and genetic counsellors. The same data were then processed using a range of different variant callers and annotation tools to evaluate their performance, and the diagnostic yield, in terms of pathogenic and likely pathogenic variants identified, was compared to conventional analysis.

Results. Alternative methods identified an additional 19 putative (likely) pathogenic single nucleotide and InDel (short insertion/deletion) variants. While the choice of SNV/InDel caller made no difference, the choice of quality cutoff or software version used for variant calling (n = 6), annotation tool used (n = 6), and new information becoming available (n = 7) impacted the diagnostic yield. For copy number variant detection, using multiple callers and a standard quality filter across them improved curation. Most new putative variants were heterozygous in genes associated with autosomal disease or found not to be phenotypically relevant to the patient. Three led to the diagnosis in unsolved trial patients, increasing the diagnostic yield from 23% (16/70) to 27% (19/70), and a further two have been referred for clinical validation.

Conclusion. This study highlights the importance of re-testing patients with no findings when new methods, technological developments and information become available.