Scholarlink

Background <p><i>Morus alba</i> is a sustainable woody biomass resource for animal feed and bioenergy production. However, its efficient preservation is hindered by high moisture levels and recalcitrant lignocellulose, which often lead to poor fermentation quality and increased greenhouse gas (GHG) emissions. In this study, we investigated an integrated strategy combining physical wilting and biochemical additives to remodel the ensiling environment and evaluated their combined effects on fermentation quality, microbial community dynamics, and GHG emissions. Furthermore, explainable machine learning was applied to decode microbial drivers.</p> Results <p>The results showed that wilting improved dry matter retention, reduced yeast and coliform bacteria, and preserved nutrients, but increased overall GHG emissions compared to the non-wilted group. The additives enhanced lactic acid production, accelerated acidification, improved fermentation quality, and reduced GHG emissions. Moreover, the combination of <i>Lactiplantibacillus plantarum</i> and cellulase maintained high levels of water-soluble carbohydrates, fostering a beneficial enzyme–microbe interaction that improved fermentation quality and reduced GHG emissions. The group supplemented with <i>L. plantarum</i> and cellulase showed the highest in vitro gas production (<i>P</i> &lt; 0.01), whereas the group supplemented with <i>L. plantarum</i> and sucrose showed the lowest in vitro gas production (<i>P</i> &lt; 0.05). Fungal communities with denitrifying potential were found to correlate with N₂O emissions. Explainable machine learning identified <i>Staphylococcus</i> and <i>Pediococcus</i> as the key predictors of GHG emissions and acidification, respectively. Furthermore, network analysis revealed that low-abundance taxa, including the bacterial hub <i>Kosakonia</i> and the fungal denitrifier <i>Paramyrothecium</i>, act as critical regulators linking fermentation quality to GHG fluxes.</p> Conclusions <p>Collectively, this study provides insights into the microbial mechanisms governing silage dynamics, confirming that the combination of wilting and the addition of <i>L. plantarum</i> and cellulase is an effective strategy for high-quality and low-carbon preservation of <i>M. alba</i> forage. Moreover, this study highlights the critical role of microbial interactions, particularly those of low-abundance taxa, in regulating fermentation quality and GHG emissions, offering a deeper understanding of the mechanisms underlying silage fermentation.</p> Graphical abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Disentangling fermentation quality and greenhouse gas fluxes in Morus alba L. silage: explainable machine learning uncovers the key microbial hubs

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月8日

查看全文
  • Ziyi Tie,
  • Xianduo Meng,
  • Cheng Chen,
  • Xiyue Niu,
  • Sasa Zuo,
  • Chuncheng Xu

摘要

Background

Morus alba is a sustainable woody biomass resource for animal feed and bioenergy production. However, its efficient preservation is hindered by high moisture levels and recalcitrant lignocellulose, which often lead to poor fermentation quality and increased greenhouse gas (GHG) emissions. In this study, we investigated an integrated strategy combining physical wilting and biochemical additives to remodel the ensiling environment and evaluated their combined effects on fermentation quality, microbial community dynamics, and GHG emissions. Furthermore, explainable machine learning was applied to decode microbial drivers.

Results

The results showed that wilting improved dry matter retention, reduced yeast and coliform bacteria, and preserved nutrients, but increased overall GHG emissions compared to the non-wilted group. The additives enhanced lactic acid production, accelerated acidification, improved fermentation quality, and reduced GHG emissions. Moreover, the combination of Lactiplantibacillus plantarum and cellulase maintained high levels of water-soluble carbohydrates, fostering a beneficial enzyme–microbe interaction that improved fermentation quality and reduced GHG emissions. The group supplemented with L. plantarum and cellulase showed the highest in vitro gas production (P < 0.01), whereas the group supplemented with L. plantarum and sucrose showed the lowest in vitro gas production (P < 0.05). Fungal communities with denitrifying potential were found to correlate with N₂O emissions. Explainable machine learning identified Staphylococcus and Pediococcus as the key predictors of GHG emissions and acidification, respectively. Furthermore, network analysis revealed that low-abundance taxa, including the bacterial hub Kosakonia and the fungal denitrifier Paramyrothecium, act as critical regulators linking fermentation quality to GHG fluxes.

Conclusions

Collectively, this study provides insights into the microbial mechanisms governing silage dynamics, confirming that the combination of wilting and the addition of L. plantarum and cellulase is an effective strategy for high-quality and low-carbon preservation of M. alba forage. Moreover, this study highlights the critical role of microbial interactions, particularly those of low-abundance taxa, in regulating fermentation quality and GHG emissions, offering a deeper understanding of the mechanisms underlying silage fermentation.

Graphical abstract