Lupus science & medicine · 2026
Balancing automation and expertise: real-world evaluation of AI-assisted whole-exome sequencing in lupus-causing genes.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Madenidou Anastasia-Vasiliki, Rice Gillian I, Garner Terence, Dyball Sarah, Parker Ben, Chieng Alice, Stevens Adam, Bruce Ian N, Briggs Tracy A
Аннотация
OBJECTIVE: Whole-exome sequencing (WES) is increasingly used to investigate patients with suspected monogenic forms of systemic lupus erythematosus (SLE) and other systemic autoimmune rheumatic diseases (SARDs). With the increasing use of artificial intelligence (AI)-assisted variant interpretation tools, there is a need to evaluate their real-world utility. In a SARD cohort, we compared the performance of AI-assisted and expert-guided WES analysis strategies in identifying disease-relevant variants within lupus-causing genes.
METHODS: We analysed WES data from 120 SARD patients using the same platform, Emedgene. For the AI-assisted analysis, a fully automated phenotype-based variant analysis produced a list of prioritised potentially disease-causing variants for each case, labelled as 'most likely'. For the expert-guided approach, expert-informed filtering criteria were applied. Both the 'most likely' AI-assisted and the expert-informed filtered variants then underwent further manual analysis.
RESULTS: Before manual analysis, a total of 2906 variants potentially likely to be causative were identified with the Emedgene AI tool (mean 20.8 variants per case), approximately three times fewer than the 8391 variants identified by the expert-guided method (mean 59.9 per case). A comparable number of variants (78 vs 56) in lupus-causing genes were identified at the end of the manual analysis with both methods, including two monogenic cases PEPD (c.819-1G>A protein change, homozygous) and CYBB (c.1314+1G>A protein change, hemizygous).
CONCLUSIONS: AI-assisted WES analysis substantially reduced the number of variants requiring manual review without missing the two monogenic cases. These findings support the integration of AI-assisted pipelines into genomic workflows for SLE and other SARDs which has the potential to improve efficiency and scalability of WES.
METHODS: We analysed WES data from 120 SARD patients using the same platform, Emedgene. For the AI-assisted analysis, a fully automated phenotype-based variant analysis produced a list of prioritised potentially disease-causing variants for each case, labelled as 'most likely'. For the expert-guided approach, expert-informed filtering criteria were applied. Both the 'most likely' AI-assisted and the expert-informed filtered variants then underwent further manual analysis.
RESULTS: Before manual analysis, a total of 2906 variants potentially likely to be causative were identified with the Emedgene AI tool (mean 20.8 variants per case), approximately three times fewer than the 8391 variants identified by the expert-guided method (mean 59.9 per case). A comparable number of variants (78 vs 56) in lupus-causing genes were identified at the end of the manual analysis with both methods, including two monogenic cases PEPD (c.819-1G>A protein change, homozygous) and CYBB (c.1314+1G>A protein change, hemizygous).
CONCLUSIONS: AI-assisted WES analysis substantially reduced the number of variants requiring manual review without missing the two monogenic cases. These findings support the integration of AI-assisted pipelines into genomic workflows for SLE and other SARDs which has the potential to improve efficiency and scalability of WES.