Skip to content
Sanatan Panda

Research

Questions I have worked on

Transcriptomics and machine learning for autoimmune disease, cancer expression and survival analysis, and computational drug safety.

IIT Bombay

RNA-seq analysis and AI models for diagnosis of autoimmune diseases

Supervisor: Prof. Prakriti Tayalia

Problem
Autoimmune diseases share overlapping symptoms and immune signatures, which complicates early and specific diagnosis. The thesis asked whether blood transcriptomes carry disease-specific signals that a classifier can learn.
Data

Public bulk RNA-seq from peripheral blood mononuclear cells (PBMCs).

  • AS23
  • RA63
  • SLE173
  • T1D137
  • Healthy43
Methods
  • DESeq2 normalization and differential expression
  • Gene set enrichment analysis (GSEA)
  • Protein-protein interaction and gene regulatory networks
  • Principal component analysis (PCA)
  • Immune cell deconvolution of PBMC expression profiles with CIBERSORT
  • GSEA run separately for each disease
  • Classifiers: Logistic regression, SVM, Ensemble model, ANN. Evaluated on an external test set.

Compute: IIT Bombay Spacetime HPC cluster (Linux).

Findings
  • Classifier features combined top-ranked GSEA genes, highly significant differentially expressed genes, and previously reported biomarkers.
  • Decision thresholds were learned from the data instead of fixed at 0.5, to minimize false negatives.
  • Models were compared on accuracy, sensitivity, specificity, F1 score, and AUC.
  • Protein-protein interaction and regulatory networks confirmed inflammatory pathways shared across the diseases, including IL/JAK-STAT, TNF-alpha, and NF-κB signaling.

M.Tech thesis result

Diagnosing autoimmune disease from blood transcriptomes

Best AUC by disease (ANN), external test set. 439 PBMC RNA-seq samples after QC.

chance 0.51.0RAn = 630.97SLEn = 1730.80T1Dn = 1370.80ASn = 230.75

Pipeline

  1. 01

    Data

    Public PBMC bulk RNA-seq

  2. 02

    QC and trimming

    Trimmomatic

  3. 03

    Alignment

    HISAT2 (hg38)

  4. 04

    Counting

    samtools, featureCounts

  5. 05

    Expression

    DESeq2, GSEA, networks, PCA

  6. 06

    Prediction

    Logistic regression, SVM, Ensemble model, ANN

Manuscript

An integrative transcriptomic approach to identify shared and disease-specific molecular signatures and predictive gene signatures across autoimmune diseases

Submitted5th authorBuilds on the M.Tech thesis

IIT Bombay, Oct 2022

LMNA (Lamin A/C) and SUN2 expression in TCGA breast cancer

Problem
Nuclear envelope proteins Lamin A/C and SUN2 are linked to nuclear mechanics. The project asked how their expression relates to clinical features and survival in breast cancer.
Data
TCGA breast cancer cohort (public).
Methods
  • Expression analysis of LMNA and SUN2
  • Survival cut-off analysis
  • Comparison across nodal stage
  • Comparison across molecular subtypes
  • Pairwise differential expression between three patient groups (C1 vs C2, C1 vs C3, C2 vs C3) on TCGA FPKM data

NIT Rourkela, May 2021

2D QSAR classification model for hERG channel inhibition

Supervisor: Prof. J Sivaraman

Problem
Blocking the hERG potassium channel can cause fatal cardiac arrhythmia, and it is a common reason drug candidates fail. The thesis built a screen to flag likely hERG blockers early.
Data
Compound structures with hERG inhibition labels, encoded as 2D molecular descriptors.
Methods
  • 2D molecular descriptors
  • QSAR classification model
  • Model selection prioritizing sensitivity and AUC over accuracy
Rationale
A missed blocker (false negative) carries a cardiotoxicity risk downstream, so sensitivity and AUC were prioritized over raw accuracy.