Diego Arturo Zavala-Trejo

Cancer remains a major global health challenge, and computational approaches can support the early-stage identification of peptide candidates for experimental evaluation. This study aimed to identify putative anticancer peptides (ACPs) encrypted within Lupinus albus β- and γ-conglutins using an in silico screening strategy. Protein sequences from two β-conglutins and two γ-conglutins were subjected to simulated enzymatic hydrolysis with pepsin, trypsin, and chymotrypsin. The resulting peptides were evaluated using PeptideRanker for bioactivity prediction, AntiCP 2.0 for anticancer peptide prediction, ToxinPred 3.0 for toxicity assessment, and ProtParam for physicochemical characterization. Simulated hydrolysis generated 1610 peptide fragments, of which 749 met the length criteria for downstream analysis. PeptideRanker identified 124 peptides with predicted bioactive potential, and AntiCP 2.0 classified 47 as putative ACPs. Toxicity screening reduced this number to 31 peptides predicted as non-toxic, resulting in 18 unique putative non-toxic ACP candidates after redundancy removal. Among these, 16 met the minimum sequence length required for complete ProtParam characterization. These findings indicate that L. albus conglutins may contain encrypted peptide sequences computationally prioritized as putative ACP candidates, whose activity, selectivity, toxicity, and safety require in vitro and/or in vivo validation.

Las proteínas, al ser hidrolizadas, liberan péptidos que pue-den tener efectos benéficos en el cuerpo humano. Gracias a la bioinformática, en la actualidad es posible hacer una pre-dicción de su potencial bioactivo usando plataformas compu-tacionales. Estas herramientas permiten simular procesos como la hidrólisis y evaluar la actividad de péptidos contra enfermedades o microorganismos. Así, los investigadores pueden ahorrar tiempo y recursos, guiando sus estudios de forma más precisa. Aunque no reemplazan los experimentos de laboratorio, los métodos in silico se han convertido en aliados clave de la ciencia actual.

Human caseins (α‐S1, β, and κ‐casein) represent a potential source of bioactive peptides with possible therapeutic applications, particularly for neonates. Due to the relevance of obtaining bioactive peptides, an in silico simulation of enzymatic hydrolysis with pepsin (pH > 2) was performed on these proteins using the PeptideCutter tool from Expasy. The resulting fragments were evaluated in silico using various bioinformatics platforms to predict biological activities such as toxicity (ToxinPred 3.0), antimicrobial (CAMP _R4 ), anticancer (AntiCP 2.0), anti‐inflammatory (PreAIP), and antioxidant (AnOxPePred 1.0) properties, as well as their overall bioactivity probability (PeptideRanker) and comparison against the BIOPEP‐UWM database. A total of 117 peptide fragments were generated: 40 from α‐S1‐casein, 48 from β‐casein, and 29 from κ‐casein. Of these, 26 peptides showed scores above 0.5 in PeptideRanker, suggesting potential bioactivity, and 17 were predicted to possess antimicrobial properties, primarily derived from α‐S1‐casein. Twelve peptides were identified as potentially toxic, with none originating from κ‐casein. Additionally, 18 peptides (16 dipeptides and 2 tripeptides) exhibited exact sequence matches with experimentally reported peptides in the BIOPEP‐UWM database. The results obtained in this study were compared with previous research, highlighting both similarities and differences in the generated sequences depending on the hydrolysis tool used. Although the results obtained through in silico analysis are encouraging, we emphasize the importance of experimentally confirming the bioactivity of these peptides. This study contributes to the understanding of the bioactive potential of peptides derived from human caseins and underscores their possible role in supporting neonatal health.