Explainable machine learning and DOE-assisted optimization reveal biochar-mediated regulation of germination and antioxidant responses in sorghum

Introduction Biochar-mediated regulation of seed germination is strongly influenced by pyrolysis conditions, precursor atmosphere, and applied concentrations. However, the combined evaluation of these factors requires modern statistical tools, and machine learning modeling can be used for optimization. Methods Different types of biochar were used to investigate the impact on temporal germination dynamics and antioxidant-related physiological responses. Experimental treatment combinations were generated using a design of experiments (DOE)-assisted optimization framework and subsequently analyzed by Taguchi signal-to-noise (S/N) ratio analysis and factorial regression analysis (FRA). Results and discussion Temporal germination analysis revealed the positive impact of high temperature on germination synchronization and speed. The Taguchi framework identified biochar concentration as the most dominant factor, whereas the FRA and Pareto analyses revealed the significance of interactive factors on germination and biochemical parameters. The temporal RF analysis showed strong associations with germination-related traits ( R 2 = 0.862–0.938), with early germination observations contributing strongly to model performance. However, a leakage-free RF analysis using only treatment variables showed limited predictive performance (mean R 2 = −0.012 to 0.195), indicating that the high performance of the temporal model partly reflected the mathematical relationship between daily germination observations and the derived indices. Biochemical factors further supported the results, and germination increased from 62.22% to 86.67%, while antioxidant activity increased and proline accumulation decreased, suggesting an improved defense system under optimized biochar treatments. Discussion Interpretable machine learning (ML) frameworks improved prediction and interpretation of complex plant responses under varying biochar conditions. These findings demonstrate that optimized biochar treatments could serve as a sustainable strategy by enhancing germination uniformity, early seedling establishment, and physiological resilience in sorghum, while further validating under greenhouse and field conditions.

Authors

Institutions

Publication Details

Journal
Frontiers in Artificial Intelligence
Published
2026-11-17
DOI
https://doi.org/10.3389/frai.2026.1925352
Primary Topic
Seed Germination and Physiology
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Explainable machine learning and DOE-assisted optimization reveal biochar-mediated regulation of germination and antioxidant responses in sorghum

Muhammad Tanveer Altaf, Ahmet Say, Farhan Aadil, Muhammad Aasım et al.
5 links
Frontiers in Artificial Intelligence
Seed Germination and Physiology
article

Explainable machine learning and DOE-assisted optimization reveal biochar-mediated regulation of germination and antioxidant responses in sorghum

Muhammad Tanveer Altaf, Ahmet Say, Farhan Aadil, Muhammad Aasım, Seyid Amjad Ali, Salabat Khan
article en

Abstract

Introduction Biochar-mediated regulation of seed germination is strongly influenced by pyrolysis conditions, precursor atmosphere, and applied concentrations. However, the combined evaluation of these factors requires modern statistical tools, and machine learning modeling can be used for optimization. Methods Different types of biochar were used to investigate the impact on temporal germination dynamics and antioxidant-related physiological responses. Experimental treatment combinations were generated using a design of experiments (DOE)-assisted optimization framework and subsequently analyzed by Taguchi signal-to-noise (S/N) ratio analysis and factorial regression analysis (FRA). Results and discussion Temporal germination analysis revealed the positive impact of high temperature on germination synchronization and speed. The Taguchi framework identified biochar concentration as the most dominant factor, whereas the FRA and Pareto analyses revealed the significance of interactive factors on germination and biochemical parameters. The temporal RF analysis showed strong associations with germination-related traits ( R 2 = 0.862–0.938), with early germination observations contributing strongly to model performance. However, a leakage-free RF analysis using only treatment variables showed limited predictive performance (mean R 2 = −0.012 to 0.195), indicating that the high performance of the temporal model partly reflected the mathematical relationship between daily germination observations and the derived indices. Biochemical factors further supported the results, and germination increased from 62.22% to 86.67%, while antioxidant activity increased and proline accumulation decreased, suggesting an improved defense system under optimized biochar treatments. Discussion Interpretable machine learning (ML) frameworks improved prediction and interpretation of complex plant responses under varying biochar conditions. These findings demonstrate that optimized biochar treatments could serve as a sustainable strategy by enhancing germination uniformity, early seedling establishment, and physiological resilience in sorghum, while further validating under greenhouse and field conditions.

Frontiers in Artificial IntelligenceVol. 9
Recep Tayyip Erdoğan University (TR), Bilkent University (TR), Arab Open University (BH), Sivas Bilim ve Teknoloji Üniversitesi (TR), Erciyes University (TR)
Zero hunger
Seed Germination and Physiology
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon