Multi-scale hierarchical regulation of starch chain length distribution and its consequential impact on starch digestibility: From molecular structure to granular degradation

Abstract

Starch chain length distribution (CLD) hierarchically modulates starch digestibility from enzyme recognition to granular degradation. This review summarizes CLD characteristics of amylose (short, medium, long chains) and amylopectin (A, B1, B2, B3 chains). The hierarchical regulatory rules are elucidated at molecular (chain flexibility, hydrogen bonding), aggregated (double helix, crystalline-amorphous regions), and granular (morphology, pores) scales. It further clarifies how distinct CLD profiles dynamically modulate the proportions of rapidly digestible starch (RDS), slowly digestible starch (SDS), and resistant starch (RS) by modulating enzyme accessibility and hydrolysis kinetics through hierarchical structures. Precise CLD regulation technologies (molecular modification, physical processing, biological regulation) and their digestibility modulation effects are summarized. Finally, key controversies (cross-scale transfer mechanisms of CLD regulation from molecular conformation to granular degradation, in vitro-in vivo discrepancy) and future directions (multi-scale characterization, precision low-glycemic index(low-GI) design) are outlined, providing a theoretical basis for functional starch foods.

1. Introduction

As a core dietary component for the majority of the global population, starch has digestibility that is directly associated with postprandial glycemic response and closely linked to the risk of metabolic diseases including obesity and Type 2 diabetes mellitus (T2DM) (Geng et al., 2026). The glycemic index (GI) is the key indicator for evaluating the glycemic regulatory effect of starchy foods. High-GI starchy foods tend to trigger pronounced blood glucose fluctuations after ingestion, making them a critical dietary risk factor for metabolic diseases (Qadir & Wani, 2022). With the global prevalence of diabetes increasing annually, the market demand for low-glycemic index (low-GI) starch-based foods has reached an annual growth rate of over 15% (Zhu et al., 2023). Accordingly, the precise regulation of starch digestibility has become a key research focus in the field of food science and nutrition.

Based on its digestibility, starch is widely classified into three categories: rapidly digestible starch (RDS), slowly digestible starch (SDS), and resistant starch (RS). The hydrolysis rate and extent of starch are regulated by its multi-scale structural features, including molecular structure, CLD, aggregated state structure, and granular morphology. Among these features, CLD serves as the key link connecting the multi-scale structures of starch across scales. As shown in Table 1, the molecular chains of starch are divided into two main categories: amylose and amylopectin. In order to distinguish and describe the different chain lengths of linear amylose, this review follows the SEC/MALLS classification convention (Masuda et al., 2012) and classified amylose into three fractions according to its degree of polymerization (DP): short-chain amylose (DP < 500), medium-chain amylose (DP 500–2500), and long-chain amylose (DP > 2500). Amylopectin is divided into distinct chain types based on DP, namely A chains (DP 6–12), B1 chains (DP 13–24), B2 chains (DP 25–36), and B3 chains (DP >36), in accordance with the classic Hizukuri classification system. These CLD characteristics directly affect the physicochemical properties and enzymatic hydrolysis behavior of starch (Almeida et al., 2020; Zhu et al., 2020).

Table 1.

Classification of Chain Length Distribution (CLD) and Degree of Polymerization (DP) Ranges for Starch Amylose and Amylopectin Fractions.

Starch Type	Chain Type Classification	Degree of Polymerization (DP) Range	Core Structural Characteristics
Amylose	Short-chain amylose	DP < 500	High molecular flexibility, prone to forming short-range double helices and V-type inclusion complexes; serves as an important precursor for resistant starch (RS3/RS5) formation after retrogradation
	Medium-chain amylose	DP500–2500	Balances molecular rigidity and aggregation capacity, participates in gel network formation after gelatinization; modulates the proportion of slowly digestible starch (SDS) through gel entrapment effects
	Long-chain amylose	DP > 2500	High molecular rigidity, readily forms intermolecular hydrogen-bond crosslinking and enhances gel thermal stability; under certain conditions, excessive chain length may restrict double-helix packing due to steric and entanglement effects
Amylopectin	A chains	DP6–12	Outermost short unbranched chains, extremely high flexibility, readily hydrolyzed by α-amylase, one of the major contributors to rapidly digestible starch (RDS), along with amorphous-region short-chain amylose and structurally disrupted starch fractions
	B1 chains	DP13–24	Intermediate chains linking A and B2 chains, characterized by moderate branching density, and modulation of enzyme cleavage site exposure
	B2 chains	DP25–36	Is prone to forming intermolecular double helices, participates in the construction of crystalline regions, and enhances structural stability
	B3 chains	DP > 36	Long-chain branches; run through crystalline and amorphous regions, enhance granular network connectivity, and reduce digestive enzyme accessibility

Recent studies on the physicochemical and functional properties of starch have incorporated systematic characterization and analysis of CLD. Numerous studies have demonstrated that CLD is strongly correlated with the digestive properties of starch (Ajala et al., 2023; Bello-Perez et al., 2021; Hamaker, 2021). Accordingly, chain length is widely recognized as one of the key determinants of starch digestive properties, acting in concert with crystallinity, granular morphology, and processing conditions. This recognition that enables elucidation of the interaction mechanism between starch and digestive enzymes at the molecular level (Li et al., 2020). In recent years, with the development and extensive application of advanced separation technologies, including gel permeation chromatography (GPC) and high-performance anion-exchange chromatography (HPAEC), research on starch CLD has deepened (Bello-Perez et al., 2021; Li et al., 2020). Relevant studies have also progressively analyzed the correlation between CLD and the physicochemical and digestive properties of starch via mathematical models (Li et al., 2020).

Considerable progress has been made in existing studies at a single scale. For instance, Gebre et al. systematically analyzed the linear correlation between the proportion of starch B2/B3 chains and resistant starch (RS) content, providing detailed data for correlation analysis between CLD and starch digestibility at the molecular scale (Gebre, Xu, et al., 2024). Shen et al. conducted research on rice starch, clarified the critical regulatory effect of granular pore structure on starch digestion behavior, and elucidated the mechanism of the starch enzymatic hydrolysis at the granular scale (Shen et al., 2026a). In the past two years, a few studies have carried out preliminary explorations on cross-scale regulation. For example, Tian et al. initially revealed the intrinsic correlation between CLD and starch crystallinity in a corn starch system, providing a new direction for elucidating the multi-scale regulatory mechanism of CLD (Tian, Huang, et al., 2025).

Although CLD plays a significant role in the factors influencing starch digestion, it is merely one component of a complex multi-factor network. Molecular size, branching architecture, crystalline polymorphism, lipids, proteins, phosphate groups, botanical origin, processing history, and food matrix structure can independently modulate digestibility.

However, overall, most existing studies are still limited to a single scale (e.g., the linear correlation between molecular chain length and RS content), and lack a systematic elucidation of the hierarchical regulatory logic spanning “molecular conformation - aggregated state structure - granular degradation”. As a result, the core mechanism underlying the regulation of starch digestibility by CLD (e.g., how chain length indirectly modulates digestive enzyme accessibility by affecting the stability of crystalline regions) has not yet been fully elucidated. The existing research findings from single-scale studies are insufficient to support the precision design of low-glycemic index (low-GI) foods. Therefore, this review systematically constructs a multi-scale hierarchical regulatory framework for CLD-mediated starch digestibility. It comprehensively elucidates the regulatory mechanisms of CLD across different chain types at the molecular, aggregated state, and granular scales, clarifies the dynamic evolution mechanism of starch digestibility dominated by CLD, and summarizes precise CLD regulation technologies and their application status. Finally, this review identifies the core controversies and future development directions of current research, aiming to provide systematic theoretical support for the development of starch-based low-GI functional foods.

Literature Search and Selection Strategy: The literature for this review was retrieved from Web of Science Core Collection, Scopus, PubMed, and Google Scholar. The search was conducted in June 2026 and covered publications from January 2010 to June 2026, with priority given to studies published after 2018. The search terms included combinations of “starch chain length distribution”, “amylose chain length”, “amylopectin fine structure”, “A/B1/B2/B3 chains”, “starch digestibility”, “RDS/SDS/RS”, “starch multi-scale structure”, “double helix”, “crystallinity”, “granular morphology”, “enzymatic modification”, “physical modification”, and “low-GI starch”. Inclusion criteria were: (1) peer-reviewed original research articles or authoritative reviews in English; (2) studies covering cereal, tuber, and legume starches to ensure botanical breadth. (3) Exclusion criteria were: conference abstracts, editorials, and non-peer-reviewed sources; studies without clear methodological description of CLD characterization; and duplicate publications. Literature quality was assessed based on sample characterization completeness, analytical method rigor, statistical reporting, and reproducibility of key findings; when conflicting results existed across studies, both viewpoints are presented and the discrepancy is discussed in Section 5.

2. Multi-scale hierarchical structural regulation of starch chain length distribution

The multi-scale structure of starch can be hierarchically decomposed into three interconnected levels based on spatial scale and structural organization: (1) the molecular conformation level (nanometer scale, DP 6–5000), which describes the chain length, flexibility, and branching pattern of individual amylose and amylopectin molecules; (2) the aggregated state structure level (sub-micrometer scale, 10–100 nm), which encompasses double-helix assembly, crystalline-amorphous lamellar organization, and short-range ordered structures formed by intermolecular interactions; and (3) the granular structure level (micrometer scale, 1–100 μm), which includes granular morphology, size, surface pores, and internal growth rings. This three-level hierarchy follows the classical structural model of starch granules (Perez & Bertoft, 2010; Wang et al., 2024) and reflects the cascade from primary molecular structure to higher-order assembled architecture. CLD serves as the foundational variable that propagates its effects upward through all three levels.

2.1. Molecular conformation level: CLD determines the fundamental properties of starch molecules

2.1.1. Amylose CLD and molecular conformation

Amylose is a linear glucan composed of α-(1–4) glycosidic linkages, and variations in its chain length directly modulate molecular flexibility and under appropriate conditions of concentration, water activity, temperature, and retrogradation history, is prone to forming short-range double helix structures that can contribute to resistant starch (RS), particularly the RS3 (retrograded) category (Delcroix et al., 2023). However, not all DP < 500 amylose behaves similarly: its ordered-structure propensity depends jointly on chain length, molecular mobility, and environmental/processing conditions, and RS encompasses multiple structural categories (RS1–RS5) that cannot be reduced to short-chain amylose double helices alone. Medium-chain amylose (DP 500–2500) balances molecular rigidity and aggregation capacity, participates in gel network formation after starch gelatinization, and modulates the proportion of slowly digestible starch (SDS). Long-chain amylose (DP > 2500) has extremely high molecular rigidity and readily forms intermolecular hydrogen-bond crosslinking, which enhances the thermal stability of starch gels; Its effect on double helix formation is non-monotonic and condition-dependent: in the early assembly stage, long chains can inhibit short-range double-helix nucleation through steric hindrance and chain entanglement; however, under prolonged retrogradation or specific hydrothermal conditions, long-chain amylose can also participate in intermolecular association, network formation, and the construction of thermally stable resistant structures (Mohammadi & Lemkul, 2026; Wang et al., 2024).

2.1.2. Amylopectin CLD and molecular conformation

Amylopectin has a highly branched structure, consisting of a backbone and external chains of α-(1–4) glycosidic linkages, with branch points formed by α-(1–6) glycosidic linkages. Distinct CLD profiles affect the degree of molecular entanglement and spatial structure of amylopectin molecules. A chains (DP 6–12), the outermost unbranched short chains in the branched structure, have extremely high molecular flexibility, are readily hydrolyzed by α-amylase, and serve as the main source of rapidly digestible starch (RDS) (Shen et al., 2026b). B1 chains (DP 13–24) are intermediate chains linking A chains and B2 chains, exhibiting a moderate degree of branching, and their relative proportion modulates enzyme cleavage site exposure on the molecular surface. B2 chains (DP 25–36) readily form intermolecular double helices with adjacent chains, participate in crystalline region formation, and enhance the structural stability of starch (Liu et al., 2026). B3 chains (DP > 36) are long-chain branches that run through multiple crystalline and amorphous regions, strengthen the internal network connectivity of starch granules, and significantly reduce digestive enzyme accessibility (Chen et al., 2026).

2.1.3. Synergistic effect of amylose and amylopectin CLD

Amylose–amylopectin chain length ratios modulate the global conformational landscape of starch via synergistic intermolecular interactions (hydrogen bonding and hydrophobic stacking). Short-chain amylose (DP < 500) coexisting with long-branched amylopectin (B2/B3) promotes double-helix assembly and crystallite nucleation via chain intercalation and alignment, whereas dominant long-chain amylose (DP > 2500) suppresses ordered packing through entanglement and hydrogen-bond disruption. Thus, the ratio bidirectionally regulates chain packing, double-helix assembly, and crystallite nucleation (Sun et al., 2024). In rice starch enriched in short-chain amylose and long-branched amylopectin, short amylose chains may intercalate within the interstitial spaces of amylopectin's branched architecture. Hypothetically, such intercalation could reinforce intermolecular hydrogen bonding between adjacent chains and help align glucan backbones into more ordered arrays, thereby promoting cooperative double-helix formation and enhancing crystalline domain development (Ang et al., 2021). This intercalation reinforces intermolecular hydrogen bonding between adjacent chains and aligns glucan backbones into ordered arrays, thereby promoting cooperative double-helix formation and enhancing crystalline domain development. Conversely, long-chain amylose possesses greater molecular rigidity and steric bulk, which drive entanglement with the amylopectin backbone. Such entanglement disrupts regular hydrogen-bonding networks, introduces structural disorder, and impedes the ordered alignment required for rapid double-helix nucleation. However, this effect is not universally suppressive: under prolonged retrogradation or high-concentration conditions, long-chain amylose can also promote intermolecular association and the formation of stable, thermally resistant aggregated structures (Martinez et al., 2018). The net outcome depends on processing history, concentration, water activity, and the amylose-to-amylopectin ratio.

However, the intermolecular interactions and synergistic effects between amylose and amylopectin remain poorly understood to date. A limited number of studies have shown that the synergistic effects between amylose and amylopectin vary significantly with the botanical source of starch. In maize starch, when the ratio of short-chain amylose (DP < 500) to B3 chains was 1:2, the double helix aggregation efficiency increased by 40% (Li et al., 2025). In contrast, in potato starch, which has a higher proportion of B2/B3 chains, a short-chain amylose to amylopectin ratio of 1:1.5 is required to achieve a significant increase in crystallinity (Chang et al., 2019; Wang, Cheng, et al., 2025).

To date, no quantitative model has been established for the correlation between the amylose-amylopectin chain length ratio and crystalline structure. Furthermore, the regulatory effects of processing treatments (e.g., gelatinization, retrogradation) on amylose-amylopectin intermolecular interactions remain unclear. Notably, the difference in this synergistic effect directly affects enzyme cleavage site exposure during starch enzymatic hydrolysis, which is a critical factor for investigating the correlation between molecular conformation and starch digestibility.

2.2. Aggregated state structure level: CLD modulates the construction of crystalline-amorphous regions

2.2.1. Effects of CLD on double helices and crystalline structure

The formation of double helices and the assembly of crystalline regions are central to the starch aggregated state structure, with CLD serving as a critical determinant (Kang et al., 2023). Except for water content, packing density, botanical biosynthesis (including isoform profiles of starch synthases and branching enzymes), and phosphate substitution, the chain length of amylopectin B2/B3 chains (DP 25–36) is positively correlated with double helix stability, these chains can form intact double helix structures, which further assemble into A-type (cereal starch) or B-type (tuber starch) crystallites. Very short external chains (approximately DP < 10) are too short to span a stable double helix and are predominantly located in amorphous regions. However, the longer segments of A chains (DP ∼10–12) and B1 chains (DP 13–24) can participate in double-helical arrangements (Eric, 2017) and are in fact major constituents of the crystalline lamellae in the classical amylopectin cluster model, where their external portions contribute to crystallite formation (Witt et al., 2012). Amylose has a broad chain length distribution, and its regulatory effect is chain length-dependent: short-chain amylose (DP < 500), under favorable retrogradation conditions, can undergo co-crystallization with the external chains of amylopectin to increase crystallinity, whereas medium- and long-chain amylose (DP ≥ 500) tends to form an amorphous network on the surface of crystalline regions to restrict excessive crystallite growth (Zhong et al., 2020). The outcome is not solely determined by DP and may vary across botanical sources and processing regimes.

2.2.2. CLD and aggregated state rearrangement during starch retrogradation

During retrogradation, CLD drives the secondary aggregation of starch chains, which directly alters starch digestibility (Li et al., 2025). During short-term retrogradation (〈12h), short-chain amylose can rapidly rearrange with amylopectin B1 and B2 chains to form partially ordered double helices. Under these conditions, such complexes are typically hydrolyzed more slowly than fully amorphous starch but less resistant than well-ordered crystallites, thereby increasing the proportion of slowly digestible starch (SDS). However, this mapping is not universal which depending on cooling rate, storage duration, water activity, and chain concentration, rapidly reassociated linear glucans can also form more ordered and stable structures that contribute to resistant starch (particularly RS3) (te Poele et al., 2020a, te Poele et al., 2020b). Thus, the digestive category of these reassociated complexes is condition-dependent rather than intrinsically fixed as SDS. In contrast, after long-term retrogradation (> 1 d), amylopectin B2/B3 chains and medium-chain amylose slowly assemble into stable crystalline regions, which promotes the formation of resistant starch type 3(RS3) (Wang et al., 2024; Wang, Chen, et al., 2021).

The retrogradation rearrangement rates of different starch types show significant differences under the regulation of CLD. For high-amylose maize starch (amylose content of 60%), after short-term retrogradation (12h) at 4 °C, the SDS proportion increased from 8% to 22%, while that of normal rice starch only increased by 5%–8% (Wu et al., 2025). Due to the high proportion of B3 chains in potato starch, a long retrogradation period (7 d) is required to form stable RS3-type crystallites, with the RS content increasing from 5% to 28%, which is much higher than the RS increment (15%) of wheat starch under the same treatment period (Wang, Huang, et al., 2025). In addition, water activity during retrogradation also exerts a regulatory effect on CLD-dominated rearrangement. The rearrangement rate of short-chain starch is highest at a water activity of 0.6, whereas at water activities below 0.3, the mobility of molecular chains is restricted, resulting in greatly reduced efficiency of double helix formation (Wang, Wang, et al., 2021).

CLD regulates the arrangement of double helices to form lamellar structures (consisting of alternating crystalline and amorphous regions), which further assemble into the microdomain structure of starch granules. Starch with a high proportion of long branched chains (B3 chains) exhibits a compact lamellar arrangement and small pores in the amorphous regions, making it difficult for enzyme molecules to penetrate. In contrast, starch with a high proportion of short branched chains (A and B1 chains) has a loose lamellar structure and a high proportion of amorphous regions, leading to enhanced digestive enzyme accessibility (Bilatu et al., 2024).

2.3. Granular structure level: CLD-dominated assembly and properties of starch granules

2.3.1. Effects of CLD on granular morphology and internal structure

The spherical or polygonal morphology, particle size, and internal pore structure of starch granules are all regulated by CLD. Most cereal starches (e.g., maize, rice) with a high proportion of A and B1 chains in amylopectin mostly present a polygonal granular morphology with large internal pores. Most tuber starches (e.g., potato) with a high proportion of amylopectin B2/B3 chains exhibit a spherical granular morphology, a dense internal structure, and a high proportion of crystalline regions (Wei et al., 2026). High-amylose starch (amylose content >50%) has a high proportion of short-chain amylose, resulting in a rough granular surface, a tendency to form internal fissures, and a more stable structure after retrogradation (Wu et al., 2014).

2.3.2. Correlation between granular surface properties and CLD

CLD modulates hydrophilicity and hydrophobicity as well as enzyme-binding sites on the granular surface by affecting the aggregation pattern of starch molecules. When short branched chains are exposed on the granular surface, they present a high hydroxyl density and strong hydrophilicity, facilitating the binding with α-amylase. In contrast, the surface structure formed by the aggregation of long branched chains and short-chain amylose has embedded hydroxyl groups and enhanced hydrophobicity, which inhibits enzyme adsorption (Brandner et al., 2022).

As shown in Fig. 1, B2 and B3 branched chains of starch aggregate to form crystalline double helix structures, which further assemble into lamellar structures with alternating crystalline and amorphous regions. Among them, starch with longer branched chains forms dense spherical granules, while starch with short branched chains forms polyhedral granules.

Fig. 1.

Schematic diagram of the multi-scale structural cascade of starch.

3. Mechanisms underlying the hierarchical regulation of starch digestibility by CLD

The previous sections have elaborated the hierarchical regulatory mechanisms of CLD on the multi-scale structure of starch, and the dynamic changes in starch digestibility directly reflect this regulation. As shown in Fig. 2, this chapter examines the entire process of starch digestion, elucidates the hierarchical regulatory pathways of CLD at each digestion stage, and summarizes the core mechanisms by which CLD affects the differentiation of RDS/SDS/RS proportions and the dynamic evolution of starch digestibility.

Fig. 2.

Schematic diagram of the starch digestion process under the hierarchical regulation of CLD.

3.1. Initial digestion stage: Regulation of enzymatic cleavage sites on the granular surface (granular level → aggregated state level)

Starch digestion begins with the binding of α-amylase to the granular surface, and CLD regulates digestive enzyme accessibility at two hierarchical levels. First, at the granular level, starch granules with a high proportion of short branched chains (A and B1 chains) have large surface pores and strong hydrophilicity, facilitating enzyme adsorption and penetration into the granule interior. In contrast, dense granules dominated by long branched chains (B2 and B3 chains) exhibit a significant surface barrier effect, with enzyme adsorption capacity reduced by 30%–50% (Sun et al., 2021). At the aggregated state level, short-chain starch (A/B1 chains and short-chain amylose) in the amorphous regions of the granular surface is preferentially hydrolyzed by enzymes, while long-chain starch (B2/B3 chains and medium- and long-chain amylose) in the crystalline regions form stable double helix structures, impeding enzymatic hydrolysis of α-(1–4) glycosidic linkages and thereby contributing to differences in the initial digestion rate (Warren et al., 2011).

3.2. Progressive digestion stage: Degradation of aggregated state structure and chain release (aggregated state level → molecular level)

As enzymatic hydrolysis proceeds, the amorphous regions on the surface of starch granules are destroyed, and the crystalline regions are exposed and gradually degraded. Variations in CLD modulate the rate and extent of this process. Specifically, short-chain-dominated starch has a high proportion of amorphous regions and loose crystalline regions; during enzymatic hydrolysis, its aggregated state structure collapses rapidly, with continuous release of short-chain starch, which is manifested as a high RDS content (digestion extent exceeding 70%). In contrast, long-chain-dominated starch has stable crystalline regions, and enzymatic first requires disruption of the double helix structure; long-chain starch is gradually hydrolyzed into medium and short chains before further degradation, resulting in high SDS and RS contents (SDS > 20%, RS > 15%) (Miao et al., 2011; Warren et al., 2011). Furthermore, terminology unified to “double-helix stability (which determines crystalline region stability)” (Pu et al., 2022).

3.3. Final digestion stage: Complete degradation of molecular chains and product distribution (molecular level)

After the complete degradation of the aggregated state structure, the released starch molecular chains are thoroughly hydrolyzed by enzymes, and CLD affects the proportion of final digestion products at this stage. Short-chain amylose (DP < 500) can be completely hydrolyzed into glucose by α-amylase, while medium- and long-chain amylose (DP > 500) has low hydrolysis efficiency due to its high molecular rigidity, a portion of which remains as RS3. A and B1 chains of amylopectin can be rapidly hydrolyzed into maltose and glucose (the main source of RDS), while B2 and B3 chains are hydrolyzed to produce medium-chain dextrins (the source of SDS), and some unhydrolyzed long-chain dextrins enter the colon for fermentation (RS5) (Li et al., 2021). The effect of CLD on digestion products is that the proportion of medium and short chains (A/B1 chains and short-chain amylose) is positively correlated with RDS content (r = 0.86–0.91), and the proportion of long chains (B2/B3 chains and medium- and long-chain amylose) is positively correlated with SDS and RS contents (r = 0.82–0.90) (Zhu et al., 2023).

3.4. Hierarchical regulatory mechanism: The transfer process of digestibility from molecular to granular scale

The hierarchical regulation of starch digestibility by CLD does not occur independently, but follows a cascading process realized through the progressive sequence of “molecular conformation → aggregated state structure → granular properties”, as summarized in Table 2. First, the molecular chain length of starch determines the formation capacity of double helix structures, which further significantly affect the structure, proportion and stability of the crystalline regions of starch. Second, the crystallinity and porosity of the granular aggregated state structure affect digestive enzyme accessibility, and thus the initial rate of digestion. Finally, the chain length distribution of molecular chains released from granule degradation during digestion can alter the substrate concentration and enzyme activity in the hydrolytic environment, and in turn regulates the subsequent degradation rate.

Table 2.

Correlation of Key Quantitative Indicators for the Hierarchical Regulation of Starch by CLD.

Regulatory Level	Key Structural Indicator	Enzymatic Hydrolysis-Related Indicator	Digestibility Indicator	Correlation (Correlation Coefficient r)
Molecular Conformation Level	Double helix formation capacity	–	Double-helix and crystalline region stability	Positive correlation (r = 0.78–0.83)
Aggregated State Structure Level	Crystallinity	Enzyme accessibility	RDS proportion	Negative correlation between crystallinity and enzyme accessibility; positive correlation between enzyme accessibility and RDS proportion (r = 0.80–0.86)
	Porosity of amorphous regions	Enzyme penetration efficiency	Initial digestion rate	Positive correlation (r = 0.78–0.83)
Granular Structure Level	Surface hydrophilicity/hydrophobicity	Enzyme adsorption capacity	Digestion initiation rate	Positive correlation between hydrophilicity and enzyme adsorption capacity (r = 0.80–0.86)
Multi-Level	Proportion of long chains (B2/B3 chains and medium- and long-chain amylose)	Enzymatic hydrolysis rate	SDS/RS proportion	Negative correlation with enzymatic hydrolysis rate; positive correlation with SDS/RS proportion (r = 0.82–0.90)
	Proportion of short chains (A/B1 chains and short-chain amylose)	Enzymatic hydrolysis rate	RDS proportion	Positive correlation (r = 0.86–0.91)

4. Precise regulation technologies for starch CLD and digestibility optimization

4.1. Molecular level regulation: Precise modulation of chain length distribution via enzymatic modification

4.1.1. Modification via tuning starch chain length

As shown in Fig. 3, amylosucrase catalyzes the polymerization of glucose monomers to elongate amylose and amylopectin chains, thereby increasing the proportion of B2/B3 chains and raising resistant starch (RS) content by 20%–40% (Zhu et al., 2023). In contrast, for starch chain shortening, maltogenic amylase hydrolyzes α-(1–4) glycosidic linkages to shorten the external chains of amylopectin, leading to increased proportion of A and B1 chains and elevated RDS content, making it suitable for improving the sensory properties of food products (Su et al., 2026; Zhu et al., 2023).

Fig. 3.

Schematic diagram of the mechanism of enzymatic modification of starch CLD at the molecular level.

In addition to maltogenic amylase, other β-amylases can also achieve targeted shortening of starch chains. They preferentially hydrolyze the terminal glycosidic linkages of the external chains of amylopectin, resulting in a more significant increase in the proportion of A chains (up to 35%). In contrast, isoamylase exhibits higher hydrolysis efficiency towards B1 chains, making it suitable for scenarios where precise regulation of the A/B1 chain ratio is required (Almeida et al., 2020).

Combined enzymatic modification shows superior performance in chain length regulation. For example, the synergistic action of maltogenic amylase and amylosucrase enables precise tuning of starch chains: maltogenic amylase acts as an exo-enzyme that selectively hydrolyzes the outer long branches of amylopectin and non-reducing ends of amylose to generate medium-short chain dextrins, while amylosucrase uses sucrose as a glucose donor to extend short-chain primers and preserves the intact core long-chain skeleton of starch granules. This optimized “long-chain backbone-medium-short chain filling” structure retains the ability of long chains to form a continuous gluten-starch network, maintaining bread's specific volume and softness, while the increased proportion of short chains enhances starch enzymatic susceptibility. When this combined enzyme-modified starch is applied to bread making, the RDS content of the bread increases by 20%, without a significant decrease in specific volume and softness (Ajala et al., 2023).

4.1.2. Modulation of branching degree

The branching degree of starch is a core indicator that affects the spatial structure of CLD as well as starch digestibility. The essence of its regulation is to directionally alter the number and distribution of α-(1–6) glycosidic linkage branch points via enzymatic modification, thereby reconstructing the chain type ratio of amylopectin to achieve precise regulation of SDS and RS contents. At present, the two most widely used modifying enzymes are debranching enzymes and branching enzymes.

As shown in Fig. 2, the first category is debranching enzymes (pullulanase/isoamylase), which hydrolyze the α-(1–6) glycosidic linkages of amylopectin to remove branch points and generate linear short-chain amylose, thus significantly increasing the RS content, with RS proportions reaching 30%–50% after combined enzyme treatment (Akassou & Groleau, 2019). The generated RS is dominated by the RS3 type, which has high structural stability and is not prone to degradation during food processing, making it a high-quality raw material for the preparation of low-glycemic index (low-GI) functional foods.

The other category is branching enzymes (e.g., AqGBE), which introduce additional α-(1–6) branch points to increase the proportion of A and B1 chains and enhance the degree of spatial entanglement of starch molecules. Although the proportion of amorphous regions increases to some extent, the dense distribution of short-chain branches restricts the penetration and hydrolysis efficiency of digestive enzymes, enabling the starch to slowly release glucose during digestion and ultimately achieving an effective enhancement of slowly digestible starch (SDS) content. Furthermore, it has favorable gelatinization properties and texture compatibility, and does not significantly alter the sensory properties and processing performance of foods. It is suitable for the modification of daily sugar-controlled staple foods such as cooked rice, bread, and steamed buns, and can improve the slow digestion properties without compromising the edibility of the products (Xia et al., 2021).

4.2. Aggregated state level regulation: Reshaping of crystalline-amorphous regions via physical modification

The aggregated state structure of starch is the core link between molecular conformation and granular properties in the multi-scale hierarchical regulation of CLD, and also a key node for the targeted regulation of starch digestibility. Physical modification technologies can act on the crystalline-amorphous regions of starch, including conventional hydrothermal treatments, as well as emerging non-thermal technologies such as microwave, ultra-high pressure, and ultrasound treatment. These technologies achieve targeted reshaping of the crystalline-amorphous region structure by regulating the aggregation behavior of starch chains. This section systematically elaborates the action mechanism, CLD regulatory mechanisms, and the impact of various technologies on starch digestibility.

4.2.1. Heat-moisture treatment (HMT)

HMT is performed at 80–140 °C with a moisture content of 15%–30%, which promotes the rearrangement of long-chain starch (B2/B3 chains and medium- and long-chain amylose), enhances the stability of crystalline regions, and can increase the resistant starch (RS) content by 10%–25% (Zhang et al., 2024). Different moisture contents lead to distinct regulatory effects on CLD, and the optimal SDS/RS ratio is achieved at a moisture content of 20% (Kumar et al., 2023).

The core regulatory mechanism of this technology is that starch granules do not undergo gelatinization in a low-moisture environment, and only limited movement of molecular chains occurs. With the aid of thermal treatment, the original loose intermolecular hydrogen bonds are disrupted,allowing long-chain starch to rearrange and form more stable double helix structures and crystalline regions. Meanwhile, the rearranged B2/B3 chains can effectively run through the crystalline and amorphous regions, further enhancing the compactness of the aggregated state structure (Zdybel et al., 2014), thereby reducing digestive enzyme accessibility.

Furthermore, the synergistic application of HMT with other modification technologies can further strengthen the targeted regulation of CLD. For example, mild HMT combined with debranching enzyme treatment can first open partial branched chain structures through thermal action, after which debranching enzymes precisely cleave the branch points. This process increases the co-crystallization efficiency of linear long-chain amylose and amylopectin B2/B3 chains by more than 20%, and the RS content increase is 10%–15% higher than that achieved with HMT alone. Owing to its simple operation and controllable cost, this synergistic process has become a commonly used technical method for optimizing starch digestibility in the food industry (Zhang et al., 2025).

4.2.2. Microwave and cold plasma treatment

Microwave treatment can break the intermolecular hydrogen bonds of starch, promote the aggregation of short-chain starch, reduce crystallinity, and lead to a short-term increase in rapidly digestible starch (RDS) content (Ma et al., 2020). The regulation of cold plasma on CLD is mainly achieved through the etching effect of high-energy particles: it preferentially breaks the outer long chains (B2/B3 chains) of amylopectin, increasing the proportion of A/B1 chains by 20%–25%, and simultaneously forming micropores with a pore size of 50–200 nm on the granular surface, which significantly enhances enzyme accessibility.

The process parameters of the microwave and cold plasma synergistic treatment have a significant impact on the regulatory effect. For example, lotus root starch treated with 700 W microwave for 5 min followed by 10 kV cold plasma for 3 min showed an increase in RS content from 8% to 35%. In contrast, if the cold plasma treatment time exceeds 5 min, the granular structure is damaged, resulting in a decreased regulatory effect on CLD (Sun et al., 2022; Sun et al., 2023).

4.2.3. Ultra-high pressure (UHP) treatment

UHP treatment is a typical non-thermal physical modification technology, which is usually performed at pressures ranging from 100 to 600 MPa and at room or mild temperatures. Its core mechanism is to reconstruct intermolecular forces through high pressure, break the original weak hydrogen bond connections between starch chains, promote water penetration into the interior of starch granules, and thus drive the differential rearrangement of starch molecules with different chain lengths and the reconstruction of double helix structures, while fully preserving the CLD profile (Wang et al., 2021).

At a treatment pressure of 300–600 MPa, high pressure can break amylopectin B2/B3 chains, which form intermolecular double helices with short-chain amylose, promote the homogeneous distribution of crystalline regions, reduce the formation of large-size amorphous regions, and significantly enhance the compactness and stability of crystalline regions (Pu et al., 2022). When the pressure is lower than 300 MPa, it can only promote the local rearrangement of A/B1 short chains in the amorphous regions and optimize the network structure of the amorphous regions, with no significant effect on the overall crystallinity. For example, after potato starch was treated with 500 MPa UHP for 15 min, the proportion of B-type crystallites formed by B2/B3 chains increased by 12%, the lamellar thickness of crystalline regions increased, and the porosity of amorphous regions decreased significantly (Cheng et al., 2021).

In terms of digestibility regulation, UHP treatment at 400–600 MPa can increase the RS content of starch to over 25%, along with a stable increase in SDS proportion. Its digestion resistance is mainly attributed to the enhanced stability of crystalline regions and the reduced enzyme accessibility in amorphous regions. In contrast, low-pressure treatment at 100–300 MPa mostly shows a slight increase in RDS content, which is due to the slight loosening of amorphous regions and the exposure of enzyme cleavage sites (Cai et al., 2022). UHP treatment can simultaneously achieve low-temperature gelatinization of starch and aggregated state regulation, avoiding the destruction of thermally unstable functional ingredients by thermal processing. It is especially suitable for the development of low-glycemic index (low-GI) starch-based food systems incorporated with active ingredients such as polyphenols and probiotics.

4.2.4. Ultrasonic treatment (UT)

Ultrasonic treatment is a non-thermal physical modification technology that achieves targeted regulation of starch aggregated state structure through the cavitation effect generated by high-frequency sound waves above 20 kHz in starch aqueous systems. It has the characteristics of high treatment efficiency, absence of chemical reagents, and strong parameter controllability, and is a widely used technical method in the field of starch green modification at present (Yao et al., 2025). Its principle is that the formation, growth and instantaneous collapse of cavitation bubbles during sound wave transmission generate local ultra-high temperature, ultra-high pressure and high-speed micro-jets, which act on the crystalline and amorphous regions of starch granules, disrupt weak hydrogen bonds between starch chains, and drive the differential movement and rearrangement of starch molecules with different chain lengths. Notably, ultrasonic cavitation shows distinct chain-length selectivity, especially for amylopectin. At low power (<300 W), it only disrupts intermolecular hydrogen bonds in amorphous regions without cleaving glycosidic bonds. At optimal power (300–600 W), it selectively breaks α-1,4 glycosidic bonds in amylopectin's outer A/B1 branches (exposed in amorphous regions), while preserving internal B2/B3 long chains and the backbone embedded in crystalline regions. This causes a slight shift in fine CLD towards shorter chains but retains starch's overall molecular weight distribution and backbone integrity It can achieve the alteration of aggregated state structure without significantly destroying the backbone glycosidic linkages of starch or changing the overall distribution characteristics of CLD (Zhang, Chen, et al., 2020).

In terms of digestibility regulation, appropriate ultrasonic treatment at 300–600 W can increase the RS content of starch by 8%–18% and the SDS proportion by 10%–15%. There are two main reasons for its digestion resistance: first, the enhanced stability of crystalline regions formed by B2/B3 chains reduces the hydrolysis efficiency of amylase; second, the structural alteration of amorphous regions forms a dense and rigid structure, reducing the accessible sites and penetration efficiency of enzyme molecules (Zhang, Dai, et al., 2020). In addition, ultrasonic treatment can be used in combination with hydrothermal treatment, enzymatic modification, pulsed electric field (PEF), and other technologies (Bai et al., 2024; Cao & Gao, 2020; Pinyo et al., 2017). Ultrasonic pretreatment opens the dense structure of starch granules, promotes the rearrangement efficiency of starch chains in the subsequent modification process, and further strengthens the aggregated state regulation effect of CLD.

4.3. Granular level regulation: Processing and breeding technologies

4.3.1. Processing technology regulation

The principle of extrusion processing is the synergistic action of shear force and heat, which breaks long-chain starch, increases the proportion of A/B1 chains, and simultaneously promotes the formation of starch-protein and starch-lipid complexes, resulting in a 15%–20% increase in RS content (Zdybel et al., 2014). Fermentation and germination usually degrade long-chain starch through microbial enzymes or endogenous enzymes. Such as the alpha-amylase-encoding gene AMY11 in banana has been identified as a key endogenous starch-degrading enzyme, whose expression drives the hydrolysis of starch chains during fruit ripening and leads to a shift of CLD towards short chains (Sun et al., 2024), leading to a shift of CLD towards short chains, while altering the granular structure and reducing the extent of starch digestion (Bui et al., 2020; Hernandez-Hernandez et al., 2019; Zdybel et al., 2014). During lactic acid bacteria fermentation, the amylase secreted by the bacteria preferentially hydrolyzes the B2/B3 chains of amylopectin, increasing the proportion of A/B1 chains in fermented rice starch from 40% to 65%. Meanwhile, starch binds to organic acids produced during fermentation to form complexes that are not readily hydrolyzed by enzymes, reducing the extent of digestion from 85% to 60% (Yuan et al., 2026).

4.3.2. Biological breeding and gene editing

Targeted alteration of CLD can be achieved by regulating the expression of starch synthase genes (e.g., SSIII, granule-bound starch synthase I (GBSSI)). For example, inhibition of the SSIII gene can increase the proportion of amylopectin B2/B3 chains and enhance RS content. Overexpression of the GBSSI gene can increase amylose content, especially the proportion of short-chain amylose, and improve the slow-digesting properties of starch (Lin, 2008; Yang et al., 2018).

4.4. Comparison of digestibility optimization effects of different regulation technologies

(Table 3)

Table 3.

Comparison of Digestibility Optimization Effects of Different Regulation Technologies.

Regulation Technology	CLD Alteration Characteristics	Digestibility Optimization Effect	Application Scenarios
Debranching Enzyme Modification	Increased linear short-chain amylose, relative enrichment of B2/B3 chains	RS content increased by 30%–50%	Low-GI functional food
Branching Enzyme Modification	Increased proportion of A/B1 chains, enhanced branching density	SDS content increased by 15%–25% (relative increase; absolute SDS content under optimal conditions: AqGBE, 5 U/g starch, 50 °C, pH 6.0, 6 h; (Xia et al., 2021))	Sugar-controlled staple foods (cooked rice, bread)
Heat-Moisture Treatment (HMT)	Rearrangement of long-chain starch, increased crystallinity	RS content increased by 10%–25%	Convenience foods (instant porridge, cereal flakes)
Microwave + Cold Plasma Treatment	Reconstruction of amylopectin CLD, increased granular porosity	RS content increased by over 30%	Functional starch raw materials
Gene Editing	Targeted enrichment of B2/B3 chains or short-chain amylose	Precise regulation of SDS/RS ratio	Specialized crop variety breeding (high-RS rice)

5. Controversies, limitations and future perspectives

5.1. Core controversies and unresolved issues

5.1.1. Controversy over the relative weight of CLD and other structural factors

Existing studies have preliminarily constructed multiple regression models to quantify the regulatory contributions of different structural factors. For instance, Liang et al. found that the relative weights of CLD, crystallinity, and granule size to starch digestion extent were 55%, 30%, and 15%, respectively (Liang et al., 2025). Whereas the study by Tian et al. showed that the weight of CLD was as high as 70%, with crystallinity accounting for only 20%, indicating that the botanical origin of starch is the core variable affecting the relative weight of each factor (Tian, Riaz, et al., 2025). In high-amylose starch, CLD directly dominates starch digestibility by regulating double helix formation; in waxy starch, the restriction effect of crystallinity on digestive enzyme accessibility is more prominent.

At present, the key difficulty in the comprehensive weight analysis of CLD lies in the significant interaction among various structural factors. For example, CLD directly regulates the formation of crystallinity, and granule size affects the contact efficiency between enzymes and starch chains with different CLD profiles, making it difficult to achieve independent analysis of a single factor.

5.1.2. Discrepancy between in vitro and in vivo digestion

The specific manifestation of the in vitro-in vivo digestion discrepancy is that, for a long-chain maize starch, the resistant starch (RS) content was 32% in an in vitro enzymatic hydrolysis experiment, while the actual undigested proportion was only 21% in an in vivo mouse experiment after fermentation by gut microbiota, representing a relative deviation of 34% (Cai et al., 2022). This deviation mainly stems from the following four aspects:

First, differences in enzyme system composition, source and activity. Current mainstream in vitro digestion models (e.g., the Englyst method, INFOGEST static model) mostly use a single enzyme system constructed from purified porcine pancreatic α-amylase and amyloglucosidase, with fixed and saturating enzyme activity, which can only simulate the basic hydrolysis process of starch (Brodkorb et al., 2019; Englyst et al., 1992). In contrast, in vivo digestion in the human body is a synergistic action of multiple enzyme systems, including salivary amylase, gastric amylase, brush border maltase-glucoamylase, and isomaltase, and the enzyme activity is affected by feeding status, individual hormone levels, and intestinal health (Li & Hu, 2023). The single α-amylase used in vitro has low hydrolysis efficiency towards amylopectin B2/B3 long chains and medium- and long-chain amylose, which tends to misclassify these long-chain fractions as RS. However, the complex enzyme system at the intestinal brush border in vivo can continuously hydrolyze long-chain dextrins, which greatly reduces the actual resistance to enzymatic hydrolysis of long-chain starch. This is the core reason why in vitro models generally overestimate the RS content of starch with a high proportion of long chains (Kim et al., 2020; Li et al., 2019).

Second, differences in digestion reaction systems. The in vitro digestion model is a closed, homogeneous system with uniform stirring, where the substrate and enzyme can achieve sufficient and unobstructed contact, and the digestion process is only limited by the enzymatic hydrolysis reaction rate. In contrast, in vivo digestion is an open continuous flow system, where the mixing pattern, retention time, and transport efficiency of food in the gastrointestinal tract are heterogeneous, and are affected by food viscosity and gastrointestinal motility (Bui et al., 2020). This difference is further amplified for starches with different CLD profiles, particularly in that starch with a high proportion of short chains (A/B1 chains) has low viscosity after gelatinization, leading to a rapid gastric emptying rate and shorter effective digestion time in the small intestine. In contrast, starch with a high proportion of B2/B3 long chains tends to form a dense gel network, resulting in higher viscosity, slower gastric emptying rate, and longer digestion time in the small intestine. Meanwhile, the gel network can encapsulate part of the starch chains and hinder the contact between enzymes and substrates (Repin et al., 2017).

Third, differences in dynamic changes of digestion environment and component integrity. The pH, temperature, and ionic strength of in vitro models are preset and constant, and most models simplify the physiological components in the gastrointestinal tract. However, the pH of the human digestive environment from the oral cavity to the colon undergoes continuous dynamic changes (oral cavity: 6.6–7.1, stomach: 1.5–3.5, duodenum: 5.5–6.5, ileum: 6.5–7.5), along with multiple influencing factors such as gastrointestinal buffer systems, bile salts, and complex dietary components (Mulet-Cabero et al., 2020; Ou et al., 2023). Among them, the strong acid environment in the stomach causes irreversible changes in the molecular conformation of B3 chains and destroys their original double helix aggregated state structure. However, most in vitro models only set a gastric digestion step with a single pH, which cannot simulate the effect of the strong acid environment on the aggregated state structure of starch with different CLD profiles (Zhuang et al., 2024). In addition, the secretion of bile salts in vivo is dynamically regulated by feeding behavior, which can promote the formation of RS5-type resistant starch by amylose and lipids. In contrast, the concentration and type of added bile salts in in vitro models are fixed, making it difficult to reproduce this process, resulting in distorted evaluation of the RS formation potential of starch with different CLD profiles (Grundy et al., 2016; Zhou et al., 2025).

Fourth, differences in digestion nodes and indicator definitions. The classification of RDS, SDS and RS in in vitro models is based on the hydrolysis rate at fixed time points (e.g., the Englyst method uses 20 min and 120 min as detection nodes). However, the physiological definition of RS is starch that cannot be hydrolyzed and absorbed in the human small intestine and can enter the colon for fermentation, and SDS is starch that is slowly and continuously hydrolyzed in the small intestine. Both are based on the continuous digestion process of the entire intestinal tract without fixed time-point restrictions (Hamaker, 2021). This ultimately leads to a deviation between the classification of starch digestion types based on different CLD profiles in vitro and the actual in vivo situation.

In addition to the above methodological differences, the regulatory effect of the intestinal microenvironment is another core contributing factor to the in vitro-in vivo digestion discrepancy. Bacteroidetes in the gut microbiota can secrete debranching amylases, which can hydrolyze B3 chain long-chain dextrins that are difficult to degrade during in vitro enzymatic hydrolysis (Butorac et al., 2026). Meanwhile, the intestinal pH (6.0–7.5) differs from the in vitro enzymatic hydrolysis pH (5.0–5.5), which can alter the conformation of starch molecules and expose enzyme cleavage sites on part of the long chains.

5.1.3. Technical limitations in precise characterization of CLD

Existing technologies make it difficult to achieve synchronous characterization from single granule to aggregated state to molecule. Size-Exclusion Chromatography (SEC) can cover the full molecular weight range (DP 10–10⁶) of both amylose and amylopectin, and when coupled with multi-angle laser light scattering (MALLS), it provides absolute molecular weight and size distribution data. Fluorophore-Assisted Carbohydrate Electrophoresis (FACE) offers exceptional single-unit resolution (can distinguish chains differing by one glucose residue), high quantitative accuracy, and ultra-low sample consumption, even enabling single starch granule analysis; yet it is limited to detecting chains with DP < 150 (Govindaraju et al., 2021). In addition to the above technologies, High-Performance Anion-Exchange Chromatography with Pulsed Amperometric Detection (HPAEC-PAD) is the gold standard for amylopectin CLD profiling due to its label-free operation, wide linear range, and excellent reproducibility, but has a slow detection speed (2–3 h per test), making it difficult to achieve rapid analysis of a large number of samples. Atomic Force Microscopy (AFM) can directly visualize the morphology, conformation, and length of individual starch molecules, providing intuitive molecular-scale structural information, but it is difficult to perform statistical analysis of CLD at the granular scale, and the detection results are significantly affected by the sample preparation method (Zhu, 2017).

At present, combinations of multiple technologies has been used to compensate for the defects of single technology. For example, the combined use of SEC and FACE can extend the chain length detection range to DP < 1000 and reduce the impact of the peak broadening effect. The combination of cryo-electron microscopy and Raman spectroscopy can achieve semi-quantitative analysis of chain length in crystalline regions while observing the microstructure. However, this combined technology has high operational complexity and expensive detection cost, making it difficult to implement widely (Shen et al., 2026a). An ideal CLD characterization technology should have four core characteristics: high resolution, rapid detection, multi-scale synchronous quantification, and non-destructive analysis, which is the core research and development direction of starch structure characterization technology in the future.

5.2. Future perspectives

5.2.1. Construction of a weight quantification system for the multi-scale hierarchical regulation of CLD

Aiming at the core controversies that the regulatory weight of CLD and other factors (crystallinity, granular structure, etc.) on starch digestibility is unclear, and the interaction between factors is difficult to decouple, future research can construct a quantitative model of multi-factor interaction by first accurately analyzing the independent contribution of a single factor to digestibility. Based on a large-sample database of starch structure and digestibility, using partial least squares regression, structural equation modeling (SEM), and machine learning algorithms, future studies can elucidate the interaction path between CLD and other structural factors, quantify the hierarchical transfer efficiency of CLD indirectly affecting digestibility by regulating aggregated state and granular structure, and clarify the contribution ratio of direct regulation by CLD and indirect regulation by downstream structures. This will improve the theoretical framework of CLD hierarchical regulation, identify the core action nodes of different chain length fractions in the cascade of “molecular conformation → aggregated state level → granular degradation”, determine the core factors leading to the differences in weight results among different studies, and provide a theoretical basis for the targeted regulation of starch digestibility by CLD.

5.2.2. Establishment of a standardized digestion method system with in vitro-in vivo correlation to achieve accurate prediction of the digestive effects of CLD

Construct a full-stage multi-enzyme synergistic in vitro digestion system. Building on classic models such as INFOGEST, supplement the complex enzyme system include human salivary amylase, gastric amylase, especially small intestinal brush border maltase-glucoamylase and sucrose-isomaltase, to match the dynamic secretion patterns and activity difference of in vivo enzymes. Optimize the enzyme system ratio for starches with different CLD characteristics (short chain A/B1 dominated, long chain B2/B3 dominated), to solve the problems of low hydrolysis efficiency of in vitro single enzyme system for long-chain starch and deviation in RS quantification.

Develop an AI-optimized dynamic gastrointestinal digestion simulation system. Based on dynamic gastrointestinal models such as TIM-1 and DGM (Lo Curto et al., 2011; Pentafragka et al., 2022), match the viscosity and gel properties of gelatinized starch with different CLD profiles, dynamically optimize gastric emptying rate, food mixing pattern and shear force parameters with the aid of reinforcement learning algorithms, restore the in vivo digestion environment, accurately distinguish the real in vivo digestion process of starch with different CLD profiles, and correct the biases in SDS and RS quantification by static models. Construct a full-scale physiological environment simulation platform with continuous pH gradient from oral cavity to colon, gastrointestinal buffer system, and dynamic bile salt secretion, to restore the irreversible remodeling effect of the strong acid environment in the stomach on the molecular conformation of long-chain amylose and B3 long branched chains. Abandon the RDS/SDS/RS classification method based on fixed time points of 20 min and 120 min, construct a continuous hydrolysis kinetic curve based on the physiological digestion duration of the human small intestine (0–360 min), replace the hydrolysis rate at fixed time points with the intestinal hydrolysis rate constant, match the correspondence between different CLD ranges and digestion kinetic characteristics, and unify the in vitro indicators with in vivo physiological definitions.

Construct a calibration and prediction model for in vitro and in vivo digestion results. Combine the standardized small intestinal digestion system with the colon fermentation system of human fecal microbiota to completely simulate the full digestion path of starch with different CLD profiles. Using a large database of in vivo and in vitro digestion results, construct a calibration model with CLD characteristics and aggregated state structure as core variables through machine learning algorithms, clarify the bias correction coefficient of starch with different chain lengths, and achieve accurate prediction of real in vivo digestion behavior through rapid in vitro detection. Explore the interaction between CLD and gut microbiota in the intestinal microenvironment (Butorac et al., 2026), and elucidate the indirect regulation of fermentation metabolites on starch digestibility. Design CLD characteristics adapted to different populations (healthy/diabetic/obese) based on their digestive enzyme activity and gut microbiota composition, to develop personalized low-GI foods.

5.2.3. Multi-scale synchronous characterization technology

Develop a full-range high-resolution detection platform for CLD, integrate the technical advantages of Size-Exclusion Chromatography (SEC), High-Performance Anion-Exchange Chromatography with Pulsed Amperometric Detection (HPAEC-PAD), and Fluorophore-Assisted Carbohydrate Electrophoresis (FACE), to construct a full-range chain length detection system for amylose and amylopectin with DP 6–5000, overcoming the detection range limitations of individual technologies. AI algorithms can be used to optimize the correction method for the peak broadening effect of SEC, improve the detection accuracy of long-chain fractions, and achieve rapid, accurate and full-range quantification of CLD.

5.2.4. Interdisciplinary extension

Strengthen the interdisciplinary integration of food science, nutrition, clinical medicine, and AI-driven big data. Clarify the causal association between starch digestibility regulated by CLD and metabolic diseases such as obesity and Type 2 diabetes mellitus (T2DM) through population study data and machine learning algorithms. Based on the characteristics of digestive enzyme activity and gut microbiota composition of different populations (healthy people, diabetic people, obese people), design starch with personalized CLD characteristics, and develop precision nutrition-oriented low-GI functional foods.

6. Conclusion

Starch chain length distribution (CLD) affects the evolution of starch digestibility through the hierarchical regulation cascade of “molecular conformation → aggregated state structure → granular degradation”. Short-chain amylose and long- branched amylopectin (B2/B3 chains) increase the SDS and RS contents by promoting double helix formation and crystalline region stability. Short-branched amylopectin (A/B1 chains) and long-chain amylose increase the RDS content by enhancing enzyme accessibility or inhibiting crystallization. Technologies such as enzymatic modification, physical processing, and gene editing can achieve targeted regulation of CLD to optimize starch digestibility.

Current research still faces challenges such as an unclear understanding of inter-scale regulatory mechanisms and the in vitro-in vivo digestion discrepancy. In the future, it is necessary to reveal the core mechanism of CLD regulating starch digestibility through the combination of multi-scale technologies, standardization of research methods, and establishment of AI intelligent models. Meanwhile, developing precise, efficient and low-cost regulation technologies will promote the large-scale production and personalized application of starch-based low-GI functional foods, and provide new strategies for the prevention and management of metabolic diseases.

CRediT authorship contribution statement

Zixuan Liu: Writing – original draft, Data curation, Conceptualization. Tianshuang Xia: Writing – review & editing, Visualization, Formal analysis. Chunfan Guo: Resources, Methodology, Investigation. Jingxuan Ma: Writing – review & editing, Supervision, Project administration. Mingyue Chen: Validation, Resources. Zhaoxia Wu: Writing – review & editing, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

No data was used for the research described in the article.