The FASTA format sequence of TINF2 protein was extracted from the UniProt database using UniProt ID (Q9BSI4). Using different databases such as dbSNP (Sherry et al., 2001), HGMD (Stenson et al., 2009), Clinvar (Landrum et al., 2014), and Ensembl (Hubbard et al., 2002) make a list of missense mutations and remove the duplicate nsSNPs. Approximately 900 mutations are obtained from the Ensembl database, of which 271 are non-synonymous mutations. The 3-D crystal structure obtained from the PDB (Berman et al., 2002) database of TINF2 protein using (PDB ID: 5XYF). This structure was used for structural-based analysis.

SIFT The Sorting Intolerant from Tolerant (SIFT) program (http://sift.jcvi.org/) uses physical properties of amino acids to predict whether amino acid substitutions are deleterious or neutral. SIFT also estimates a protein’s sequence homology. If the SIFT score is more than 0.05, then the mutation is tolerable or neutral; if the score is equal to or less than 0.05, then it is not tolerable or deleterious (Kumar et al., 2009). All 271 nsSNPs mutations of the TINF2 protein are tested using the SIFT tool.

PolyPhen-2 Polymorphism phenotyping (PolyPhen-2) (http://genetics.bwh.harvard.edu/pph2/) is a sequence-based tool that finds the damaging probability of amino acid substitution in any protein on the basis of its physical and comparative properties (Ramensky et al., 2002). It takes a FASTA format sequence as input and estimates the PSIC score between 0 to 1. PolyPhen-2 scores range from 0.0 to 1.0. Variants closer to 1.0 are more likely to be damaging and are classified as benign, possibly damaging, or probably damaging based on predefined thresholds.

PROVEAN Protein variation effect analyzer (PROVEAN) (http://provean.jcvi.org/) is a sequence-based tool take FASTA format sequence as an input and predicts the damaging nsSNP of protein on the basis of the functionality of the protein (Choi & Chan, 2015). If the PROVEAN score is greater than -2.5, the mutation is predicted to be neutral; scores less than -2.5 are considered deleterious.

Mutation Assessor Mutation Assessor (http://mutationassessor.org/r3/) is a sequence-based tool that takes a UniProt ID or NCBI RefSeq ID as input for a protein. The mutation assessor classifies mutations as medium, low, or neutral in terms of deleteriousness based on multiple sequence alignment and evolutionary conservation (Reva et al., 2011). This tool predicts the functional impact of a missense mutation on a protein and computes the FI score. If the FI score is less than 2.00, the mutation is neutral; if the FI score is greater than 2.00, the mutation is considered deleterious.

SAAFEC-SEQ Single Amino Acid Folding free Energy Changes (SAAFEC-SEQ) (http://compbio.clemson.edu/SAAFEC-SEQ/index.php). SAAFSEC-SEQ is a sequence-based tool used to determine the change in folding free energy caused by amino acid substitution (Getov et al., 2016). It takes a protein FASTA sequence as input and predicts changes in free energy based on physicochemical properties and sequence features.

SDM2 Site-Directed Mutator (SDM2) (http://marid.bioc.cam.ac.uk/sdm2) is structure based tool that takes a PDB ID or a PDB file as input and uses environment-specific amino acid substitution tables to calculate the stability of an nsSNP in a protein (Pandurangan et al., 2017). It estimates the change in stability between the wild-type and mutant nsSNP of a protein using ΔΔ G. SDM2 estimates the change in stability (ΔΔ G) between the wild-type and mutant protein. Negative ΔΔ G values indicate destabilizing mutations, whereas positive values indicate stabilizing effects. All 134 nsSNPs located within the resolved crystal structure were analyzed using SDM2.

mCSM It is a structure-based tool to estimate nsSNP. mCSM take PDB file or a FASTA sequence as input and predict the mutation is destabilizing or not on the basis of graph based approach and calculates ΔΔ G (http://structure.bioc.cam.ac.uk/mcsm) (Pires et al., 2013). For a range of proteins, mCSM provides better comparisons of mutations linked to disease. If the score is less than 0, the mutation reverses its effect.

DUET DUET (http://biosig.unimelb.edu.au/duet/) is a structure-based tool that takes a PDB file and an nsSNP mutation as input and calculates the DUET score, along with mCSM and SDM scores, as outputs. The DUET tool employs SVM and combines the results of mCSM and SDM to compute ΔΔ G. DUET integrates residue-level RSA (BY SDM) and secondary-structure and pharmacophore (by mCSM) features, and combines them with a supervised learning program (Pires et al., 2014).

MUpro MUpro (http://mupro.proteomics.ics.uci.edu/mutation_intro.html/) is a machine-learning technique used to determine how a single amino acid mutation affects protein stability. It employs two distinct machine-learning models to compute energy change. These programs are SVM and neural networks (Cheng et al., 2006). It computes a confidence score between -1 and 1 and the change in energy to quantify confidence. If the MUpro score is greater than 0, the mutation is predicted to increase protein stability; if the score is less than 0, it is predicted to decrease protein stability.

Pmut PMut (http://mmb.irbbarcelona.org/PMut) is structure based tools which is used to check whether a non-synonymous mutation is related to disease or not (Lopez-Ferrando et al., 2017). Updated version of pmut determine. Pathogenicity of mutations, but it also provides an opportunity to train one's own prediction model on a particular dataset.

Phd-SNP Phd-SNP (http://gpcr.biocomp.unibo.it/cgi/predictors/PhD-SNP/PhD-SNP.cgi/) is a sequence or structure-based tool that is used to differentiate disease-related mutations regulated by the local environment of substitution. It takes a PDB file or a FASTA sequence as input.

RHAPSODY Rhapsody (http://rhapsody.csb.pitt.edu/) is a computational tool primarily used to determine the Pathogenicity of missense mutations based on sequence- and structure-based properties of the mutant protein (Ponzoni et al., 2020). It outperforms EV mutation and Polyphen-2 in analyzing residue-average pathogenicity.

The updated version of sdm2 also calculates the RSA, OSP, and residue depth for both mutant and wild-type proteins. For RSA, OSP, and residue-depth calculations, an environment-specific amino acid table is employed. RSA can calculate by the Lee and Richard technique. For the Structure stability analysis, OSP and residue depth are extended properties of the protein (DeDecker et al., 1996; Richards & Lim, 1993).

Arpeggio Arpeggio server is an structure based tool that takes a PDB ID or a PDB format file as input and estimates all the interatomic interactions of a protein. The Arpeggio server estimated mainly 15 different types of interatomic interactions, which can also be seen by this server (http://structure.bioc.cam.ac.uk/arpeggio/) (Jubb et al., 2017).

SODA Solubility based on Disorder and Aggregation (SODA) (http://protein.bio.unipd.it/soda/) is a sequence or structure-based tool used to estimate the aggregation of protein, disorder, helix, and strand propensity that emerge because of mutations. Use the FASTA protein sequence or the PDB structure file as input. Insertion, deletion, substitution, and duplication (mutation type) were estimated by soda using PASTA 2.0, ESpritz-NMR, and Fells Soda score estimated according to the solubility difference between the mutant and WT of protein (Paladin et al., 2017).

A list of 900 mutations was obtained from Ensembl, of which only 271 were non-synonymous. Out of 271 nsSNPs, only the structure of the N-terminal region (residue 2-202) is found in the PDB structure (PDB ID: 5XYF) of the TIN2 protein in the RCSB database. This analysis focuses on the sequence- and structure-based study of all nsSNPs. A multilevel approach was used to determine the functional and structural importance of the mutation on the TINF2 protein. To determine high-confidence deleterious and destabilizing non-synonymous mutations, sequence- and structure-based analyses were performed. All 271 non-synonymous mutations were subjected to sequence-based analysis using five tools: SIFT, PROVEAN, PolyPhen-2, Mutation Assessor, and SAAFEC-SEQ. Structure-based analysis is done by four different web-based tools, which are SDM2, DUET, mCSM, and MUpro, to estimate the 134 non-synonymous mutations present within the resolved crystal structure of the TINF2 protein (PDB ID 5XYF). For further studies, only high-confidence nsSNPs were included. Pathological analysis of high-confidence nsSNPs was performed using two web tools: PMut and PhD SNP.

All 271 nsSNPs were analyzed using five independent sequence-based prediction tools. The proportion of variants predicted as deleterious varied among tools, reflecting differences in underlying algorithms. SIFT predicted 142 variants (52.3%) as damaging, PolyPhen-2 identified 96 (35.4%) as possibly or probably damaging, PROVEAN classified 41 (15.1%) as deleterious, Mutation Assessor detected 80 (29.5%) as functionally impactful, and SAAFEC-SEQ predicted 264 (96.3%) variants to reduce folding free energy (Table S1). Although individual tools yielded variable results, consensus analysis reduced the risk of false positives (Figure). Variants predicted as deleterious by at least four of the five tools were considered high-confidence functionally damaging mutations. This approach improved reliability by integrating evolutionary conservation, physicochemical changes, and predicted changes in folding energy. The high proportion of predicted deleterious variants suggests that many substitutions in TIN2 may disrupt its structural integrity or protein–protein interactions within the shelterin complex.

Structure-based prediction was performed on 134 nsSNPs located within the resolved N-terminal structure of TIN2. Stability changes were assessed using SDM2, DUET, mCSM, and MUpro. Among the analyzed variants, SDM2 predicted 58 (43.2%) as destabilizing, DUET identified 103 (76.8%), mCSM detected 120 (89.5%), and MUpro predicted 128 (95.5%) variants as decreasing protein stability (Figure). The predominance of destabilizing predictions indicates that many substitutions in the N-terminal domain may impair structural stability. Destabilization of TIN2 is particularly significant because the N-terminal region mediates high-affinity binding with TRF2 and contributes to scaffold formation within the shelterin complex. Structural perturbations in this domain may weaken intermolecular interactions, leading to defective telomere protection and compromised telomere length regulation. To increase prediction accuracy, only variants classified as destabilizing by at least three of four structure-based tools and as deleterious by at least four of five sequence-based tools were retained. This stringent filtering yielded 35 high-confidence deleterious and destabilizing nsSNPs.

The 35 high-confidence variants were further evaluated for disease association using PMut, PhD-SNP, and Rhapsody. PMut predicted 25 (71.4%) variants as pathogenic, PhD-SNP identified 24 (68.5%), and Rhapsody classified 11 (31.4%) as disease-associated. Twelve mutations, A16D, V22G, V34G, L38P, V49G, R52P, R56G, E109G, F114S, L121P, Y139C, and L150P, were consistently predicted as pathogenic by at least two of the three tools. These variants were selected for detailed structural investigation. Interestingly, several of these mutations are located in highly conserved residues within the N-terminal region, supporting their functional importance. Many involve substitution of small or hydrophobic residues with charged or bulky amino acids, which may significantly alter local structural packing.

Changes in Stability and Packing Density Analysis of relative solvent accessibility (RSA), occluded surface packing (OSP), and residue depth revealed that most of the selected mutations induced noticeable alterations in local packing density. Mutations such as V22G, V34G, V49G, and L150P replace hydrophobic residues with glycine or proline, thereby disrupting secondary-structure elements and reducing hydrophobic-core stability. Proline substitutions (L38P, L121P, L150P) are particularly disruptive due to conformational rigidity and their tendency to break α-helices. These changes likely perturb TIN2's folding and structural integrity (Table).

Alterations in Noncovalent Interactions Interatomic interaction analysis using Arpeggio demonstrated that several pathogenic mutations reduced hydrogen bonds, hydrophobic contacts, and ionic interactions compared to the wild-type structure (Table). Loss or rearrangement of these interactions can destabilize the local microenvironment and impair protein–protein interaction interfaces. For example, substitutions involving charged residues (R52P, R56G, E109G) may abolish electrostatic interactions critically for maintaining tertiary structure or mediating binding with TRF1 and TRF2.

Aggregation Propensity and Solubility Changes Aggregation analysis using SODA indicated that nine of the twelve selected mutations showed increased aggregation propensity, whereas the remaining three showed reduced solubility. Increased aggregation propensity suggests a higher likelihood of misfolding and potential formation of non-functional protein assemblies (Table). Protein misfolding or aggregation of TIN2 could disrupt shelterin complex assembly, thereby impairing telomere capping and exposing chromosome ends to DNA damage response pathways.