IEEE Genomic Survey Paper

Datatypes and Methods

Sachin Kumar

Project BL2403, SCoS & SBS, NISER

Outline

  • Data types covered in the paper

  • Methods based on the datatypes

  • Applications

  • Comparison of Models

Not covered: Detailed discussion of models' architecture

Datatypes at different molecular level

  • Genomics Data

  • Transcriptomics Data

  • Epigenomics Data

  • Pharmacogenomics Data

  • Proteomics Data

Tasks

Prognostic Predictions

  • Detect various expressions in omics data.

  • Deeply mine complex genomics data Classify critical regulators.

  • Detect representations in every omics data.

  • Identify cancer cells with related Epigenomics.

Cancer Applications

  • Patient Survival Rate

  • Early Prognosis

  • Cancer Classification

Tasks

Therapeutics Predictions

  • Identify drugs giving a desired gene expression.

  • Drug response prediction based on mutation.

  • Sensitivity classification of each anticancer drug.

  • Detect therapeutic biomarkers for drug response prediction.

Cancer Applications

  • Drug Response

  • Drug Target Discovery

  • Biomarker Discovery

Base Deep Learning models covered

Auto Encoder, Vision Transformers, GNN, CNN, GCN, bi-LSTM, Denoising AE, SGCN, Kernel DNN, NN Triplet Loss, Multi-Task NN, CNN-LSTM, VNN

Specific Deep Learning models covered

CITRUS, RDAClone, dpAE, DeepDEP, MOMA, GTN, MOGONET, MultiCoFusion, SWnet, GTN, GCN-CRISPR, MultiCoFusion, CDNN, DeepCellEss, DeepSynergy, DeepDRK, Super.FELT, CSynergy, CCSynergy, DeepSynergy, Apindel, DeepDep, BioVNN

MOMA 13 April 2023

Multi-Omics Multi-cohort Assessment

MOMA 13 April 2023

Multi-Omics Multi-cohort Assessment

CITRUS 28 Oct 2022

Chromatin-informed Inference of Transcriptional Regulators Using Self-attention

CITRUS 28 Oct 2022

Chromatin-informed Inference of Transcriptional Regulators Using Self-attention

Overview of CITRUS: An attention-based model with TF-target gene priors. The input to our framework includes somatic alteration and copy number variation, assay for transposase-accessible chromatin with high-throughput sequencing (ATAC-seq), tumor expression datasets and TF recognition motifs. CITRUS takes somatic alteration and copy number variation data as input and encodes them as a tumor embedding using a self-attention mechanism. Cancer type embedding is used to stratify the confounding factor of tissue type. The middle layer further transforms the tumor embeddings into a TF layer, which represents the inferred activities of 320 TFs. Finally, gene expression levels are predicted from the TF activities through a TF-target gene priors constrained sparse layer based on ATAC-seq.

MultiCoFusion April 2022

multi-modal fusion framework based on multi-task correlation

code

site

MultiCoFusion April 2022

multi-modal fusion framework based on multi-task correlation

code

site

MOGONET 08 June 2021

Multi-Omics Graph cOnvolutional NETworks: patient  classification and biomarker identification

MOGONET 08 June 2021

Multi-Omics Graph cOnvolutional NETworks: patient  classification and biomarker identification

MOGONET combines GCN for multi-omics-specific learning and VCDN for multi-omics integration. For clear and concise illustration, an example of one sample is chosen to demonstrate the VCDN component for multi-omics integration. Preprocessing is first performed on each omics data type to remove noise and redundant features. Each omics-specific GCN is trained to perform class prediction using omics features and the corresponding sample similarity network generated from the omics data. The cross-omics discovery tensor is calculated from the initial predictions of omics-specific GCNs and forwarded to VCDN for final prediction. MOGONET is an end-to-end model and all networks are trained jointly.

MultiCoFusion April 2022

Chromatin-informed Inference of Transcriptional Regulators Using Self-attention

DeepPathNet 22 April 2024