#20 · Tools & infrastructure
Packaged tools post-process pathway, m6A, radiation transcriptomic, and cell death data
The dominant thread is software that packages existing resources and statistical post-processing into reusable workflows for omics and pathway data. WayFindR converts WikiPathways and KEGG pathways into igraph graphs to detect negative feedback loops in biological pathways. m6AnetAnalyzer post-processes site-level m6Anet predictions, summarizing distributions, annotating sites, and quantifying differential methylation. DoReMiTra unifies curated radiation transcriptomics microarray and RNA-seq datasets under harmonized metadata for orchestrated analysis. RCDRank prioritizes 18 regulated cell death modalities by combining seven gene set enrichment tests via Tippett p-value and robust rank aggregation. A withdrawn microbiome random-effects network preprint sits at the edge of this set, and its biological validation of inferred directions remains incomplete. What distinguishes these records is their focus on interpretation layers built atop existing detection methods or curated resources, rather than new primary assays.
Published papers
GraFSyn: An Interpretable Deep Learning Framework for Anticancer Drug Synergy via Graphlet Fingerprints.
GraFSyn uses graphlet fingerprints and a dynamic multi-scale convolution module to predict anticancer drug synergy, with interpretable substructure-level…
Unifying pKa and Protonation Prediction with Sequence-Based Deep Learning.
Introduces pKaCHU, a curated set of 9000 microscopic pKa entries, and T5pKa, a T5Chem-based sequence-to-sequence model unifying microstate enumeration and pKa…
Preprints
Mixing Vector Model for Copolymer Inference via Mixed Integer Linear Programming
Proposes a mixing vector representation enabling MILP-based inverse design of copolymers, with neural, regression, and forest predictors reaching test R^2…
GraphTox: A Semi-Supervised Pre-Trained Framework for Peptide Toxicity Prediction using Geometric Graph Transformer and LORA-Based Finetuning
GraphTox: a structure-aware geometric graph transformer with masked graph autoencoder pretraining and U-Net fusion predicts peptide toxicity from 3D structural…
Pathogen-specific antimicrobial activity prediction with biological large language model-based methods
triAMPh, a zero-shot framework combining a heterogeneous graph attention link predictor, XGBoost, and an MLP on biological LLM features, predicts…
DMPKformer: An Interpretable Multimodal Deep Learning Framework for Reliable ADMET Property Prediction
Multimodal deep learning framework (Transformer on MACCS fingerprints, GAT on graphs, MLP on descriptors) for ADMET prediction with latent OOD confidence…
Topic history
| Date | Rank |
|---|---|
| #20 | |
| #10 | |
| #8 | |
| #8 | |
| #6 | |
| #13 | |
| #8 | |
| #13 | |
| #7 | |
| #12 | |
| #22 | |
| #10 | |
| #12 | |
| #2 | |
| #38 | |
| #40 | |
| #32 | |
| #25 | |
| #36 | |
| #35 | |
| #37 | |
| #3 | |
| #33 | |
| #24 | |
| #39 | |
| #33 | |
| #31 | |
| #23 |