Biology subjects

Babu, D.

Publications and source records attributed to Babu, D..

5 recordsLinked to original sources

bioRxiv2026

Sequence-only prediction of VHH nanobody epitopes and paratopes on HER2 with a frozen protein language model ( ESM2) and a structure-derived residue prior

Mapping which antigen residues an antibody contacts (the epitope) and which antibody residues do the contacting (the paratope) is central to therapeutic antibody engineering, yet most accurate methods require a solved or predicted three dimensional complex. We address a sequence only formulation: predict the HER2 epitope and the single domain antibody (VHH, or nanobody) paratope from the two sequences alone, with no three dimensional structure prediction at any stage. We pair a frozen ESM2 650M protein language model, used strictly as a feature extractor, with a lightweight trainable module per chain encoders, a shared cross attention block, and epitope, paratope and binder heads (12.2M trainable parameters) trained jointly on three public sources: HER2 binding and blocking competition assays ("AVIDa hHER2" to be published by organizers of SEPIQ 2026 The authors have no right to share the dataset to public ), residue level antibody antigen contacts (SAbDab2), and two reference antibody crystal structures. The blocking competition signal is de confounded into a trastuzumab minus pertuzumab contrast whose sign encodes the HER2 domain and which is used only as a training target, never as a model input. Under a three way, sequence cluster level split with a test set opened exactly once, the binder head reaches PR AUC 0.999 and the paratope head 0.659, while residue level epitope prediction on novel antigens remains near baseline (0.070) a limitation the sealed split exposes rather than hides. Sequence alone recovers the competition derived HER2 domain (the competitor class) for 87.7% of held out clones, a leakage free result. For HER2 specifically, a residue prior built from 25 mined HER2 antibody complexes and validated by leave one out reaches 0.26 to 0.29 residue PR AUC once the domain is correct, roughly 3.6 times the language model head alone. We release and report two submission variants per test clone, a fully sequence only prediction and a competition informed one ,and document which data were used where. Keywords: epitope prediction; paratope prediction; nanobody; VHH; HER2; protein language model; ESM2; cross attention; antibody antigen interface; leakage safe evaluation.

bioinformatics↗

bioRxiv2023

Inhibition of BCR::ABL1 tyrosine kinase activity Aids in the Generation of Stable Chronic Myeloid Leukemia Induced Pluripotent Stem Cells

Induced pluripotent stem cells (iPSCs) generated from patients with chronic myeloid leukemia (CML) have the potential for disease modeling to study disease pathogenesis and screening therapeutic interventions. In this study, we aimed to generate iPSCs from CD34+ hematopoietic progenitors of CML patients with varying responses to tyrosine kinase inhibitor (TKI) therapy. The generated CML-CD34-iPSC colonies displayed atypical "dome-shaped" morphology and underwent spontaneous differentiation in a few days. However, supplementation with imatinib (IM), the most widely used TKI to treat CML patients, in the culture medium improved the stability and maintenance of all isolated CML-CD34-iPSC colonies, allowing them to be maintained for more than 20 passages without significant differentiation. In contrast to previous studies, our results indicate that suppressing the BCR::ABL1 oncogenic pathway is essential for efficiently generating stable CML-iPSC colonies. Furthermore, we successfully differentiated these iPSCs to CD34+ hematopoietic progenitors both in the presence and absence of IM. This robust protocol for generating CML-iPSCs provides a valuable resource for disease modelling. The generated iPSCs will be a valuable tool for investigating CML pathophysiology, drug resistance mechanisms, and drug screening to identify novel and effective therapies for this disease.

cancer biology↗

bioRxiv2021

SYS-Mut: Decoding the Functional Significance of Rare Somatic Mutations in Cancer

Current tailored-therapy efforts in cancer are largely focused on a small number of highly recurrently-mutated driver genes but therapeutic targeting of these oncogenes remains challenging. On the other hand, the vast number of genes mutated infrequently across cancers have received less attention, in part, due to a lack of understanding of their biologic significance. Here we present SYS-Mut, a systems biology platform that can robustly infer the biologic consequences of somatic mutations by integrating routine multi-omic profiles in primary tumors. We established the accuracy of SYS-Mut by recapitulating the functional impact of known driver genes in PanCancer datasets. Subsequent application of SYS-Mut on low-frequency gene mutations in Head and Neck Cancers (HNSC), followed by molecular and pharmacogenetic validation, revealed the lipidogenic network as a novel therapeutic vulnerability in aggressive HNSC. SYS-Mut is thus a robust scalable framework that enables discovery of new targetable avenues in cancer.

cancer biology↗

bioRxiv2021

The EphB2-MYC Axis is a Major Determinant of Barrett's Pathobiology and a Therapeutic Vulnerability in Esophageal Cancer

Esophageal adenocarcinoma (EAC), a highly aggressive cancer with limited therapeutic options, often arises in the backdrop of a molecularly-complex esophageal metaplasia disorder, Barretts Esophagus (BE). Using transcriptomics and systems biology analyses of treatment-naive malignant/pre-malignant biopsy tissues, we found Eph receptor B2 (EphB2) tyrosine kinase signaling to be frequently hyperactivated during early stages of EAC development, and across the BE-EAC continuum. Functional studies revealed EphB2 to be an upstream post-translational regulator of c-MYC activity and as a key molecular dependency in BE/EAC. Single-cell transcriptomics in a porcine esophageal 3D spheroid model showed enhanced EphB2 and MYC activity to be significantly associated with BE-like cell fate. shRNA-based knockdown of EphB2 or small molecule inhibitors of MEK, that modulate MYC protein stability, proved effective in suppressing EAC tumor growth in vivo. These findings point to EphB2-MYC axis as an early promoter of EAC and a novel therapeutic vulnerability in this increasingly-prevalent esophageal malignancy. STATEMENT OF SIGNIFICANCEWe identify EphB2 signaling as a potential master regulator and early promoter of esophageal adenocarcinoma, and the proto-oncogene MYC as a key downstream effector of EphB2 function. Targeting the EphB2-MYC axis could be a promising therapeutic strategy for these often refractory and lethal EAC tumors.

cancer biology↗

bioRxiv2019

Cooperation only for high rewards - a solvable task-based study on free-ranging dogs

The benefits of group living mostly surpass the disadvantages like sharing of resources and competition over food, space and mates, driving the evolution of social organization. Group living can be facilitated by social tolerance and cooperation among the group members. Social canids (e.g. wolves) display cooperative breeding, hunting, and prosocial activities in different contexts. Unlike cooperative pack-living wolves (Canis lupus lupus), their descendants, domesticated dogs (Canis lupus familiaris), show varying levels of associations from solitary to stable social groups. Free-ranging dogs are group-living but prefer to forage solitarily, hence providing an excellent opportunity for investigating social tolerance and coordinated task performance among the members in various situations. We tested 113 adult-only groups of free-ranging dogs in three different tasks to investigate group responses and performance in problem-solving situations in the presence of an unfamiliar human. Task 1 (unfamiliar, single food reward) and 2 (familiar, single food reward) examined group responses and cooperation from the perspective of familiarity, while Task 3 (familiar, multiple food rewards) enabled us to test whether increased food rewards promote social tolerance and food sharing among the group members. Regardless of significantly higher performance in Task 2 compared to Task 1, cooperation and food sharing were significantly lower in both. Task 3 revealed a strong positive correlation between food sharing and social tolerance, but not between success and social tolerance, suggesting a tendency for cooperation. We conclude that context-dependent cooperation and tolerance among group members facilitate group-living in free-ranging dogs.\n\nSignificance statementGroup living is a common phenomenon in the animal world where the members of a group show social tolerance and co-operative behaviours towards each other. This need for cooperative intents increases manifolds while groups face different problem-solving situations in their day to day lives. Here, we tested a large number of free-ranging dog groups to understand general cooperative intents such as social tolerance and food sharing in different problem-solving conditions. We found shreds of evidence of context-dependent cooperation and social tolerance among group members with minimal display of aggression. It is not adaptive for the dogs to fight or display aggression over resources. Alternatively, use of subtle cues such as display of dominance and subordination seem to be more plausible mechanisms for the development of efficient scavenging strategies and maintaining hierarchy.

animal behavior and cognition↗