Taxonomy · BioBench Atlas

Scientific domain

ID	English	中文	Parent	Definition
`life-science`	Life science	生命科学	—	Broad applied and foundational life-science research.
`protein-science`	Protein science	蛋白质科学	life-science	Protein sequence
`protein-sequence`	Protein sequence	蛋白质序列	protein-science	Amino-acid sequence understanding and modeling.
`protein-structure`	Protein structure	蛋白质结构	protein-science	Protein folding
`protein-design`	Protein design	蛋白质设计	protein-science	Designing or optimizing protein sequences
`protein-protein-binding`	Protein-protein binding	蛋白-蛋白结合	protein-science	Protein-protein interaction
`protein-ligand-binding`	Protein-ligand binding	蛋白-配体结合	protein-science	Binding between proteins and small molecules or other ligands.
`antibody-antigen`	Antibody-antigen	抗体-抗原	protein-protein-binding	Antibody specificity
`rna-protein-binding`	RNA-protein binding	RNA-蛋白结合	life-science	Interaction
`genomics`	Genomics	基因组学	life-science	Genome variation
`transcriptomics`	Transcriptomics	转录组学	life-science	Bulk and other transcriptome-level measurements and analyses.
`epigenomics`	Epigenomics	表观组学	life-science	DNA methylation
`single-cell`	Single-cell	单细胞组学	life-science	Single-cell molecular measurements and analyses.
`spatial-omics`	Spatial omics	空间组学	life-science	Spatially resolved molecular measurements and analyses.
`proteomics`	Proteomics	蛋白质组学	life-science	Proteome-scale measurement and analysis.
`metabolomics`	Metabolomics	代谢组学	life-science	Metabolite-scale measurement and analysis.
`microbiome`	Microbiome	微生物组	life-science	Microbial community composition and function.
`multiomics`	Multi-omics	多组学	life-science	Integrated analysis across two or more omics modalities.
`molecular-cell-biology`	Molecular and cell biology	分子与细胞生物学	life-science	Cellular mechanisms
`assay-screening`	Assays and screening	实验与筛选	life-science	Assay design
`bioinformatics`	Bioinformatics	生物信息学	life-science	Computational analysis of biological data and databases.
`medchem`	Medicinal chemistry	药物化学	life-science	Small molecules
`clinical-translational`	Clinical and translational	临床与转化	life-science	Preclinical-to-clinical reasoning
`virology`	Virology	病毒学	life-science	Viral sequences

Capability

ID	English	中文	Parent	Definition
`knowledge`	Knowledge	知识	—	Recall and apply scientific knowledge.
`evidence-synthesis`	Evidence synthesis	证据综合	—	Interpret and reconcile papers
`retrieval`	Retrieval	检索	—	Retrieve scientific records
`prediction`	Prediction	预测	—	Predict scientific properties or outcomes.
`classification`	Classification	分类	prediction	Assign discrete biological labels.
`regression`	Regression	回归	prediction	Predict quantitative biological values.
`design`	Design	设计	—	Design molecules
`generation`	Generation	生成	design	Generate sequences
`optimization`	Optimization	优化	design	Improve a scientific object or process under constraints.
`data-analysis`	Data analysis	数据分析	—	Analyze quantitative
`coding`	Coding	编程	data-analysis	Write and execute analysis code.
`tool-use`	Tool use	工具使用	—	Use scientific software
`experiment-planning`	Experiment planning	实验规划	design	Plan experiments
`troubleshooting`	Troubleshooting	故障排查	—	Diagnose experimental or analytical failures.
`scientific-reasoning`	Scientific reasoning	科学推理	—	Explain mechanisms
`scientific-communication`	Scientific communication	科学交流	—	Produce expert-useful scientific explanations and reports.

Input modality

ID	English	中文	Parent	Definition
`text`	Text	文本	—	Natural-language task context.
`paper`	Paper or document	论文或文档	text	Scientific papers
`table`	Table	表格	—	Tabular or spreadsheet data.
`figure`	Figure	科研图表	—	Plots
`dna-rna-sequence`	DNA or RNA sequence	DNA/RNA序列	—	Nucleotide sequences or sequence files.
`protein-sequence`	Protein sequence	蛋白质序列	—	Amino-acid sequences or sequence files.
`small-molecule-structure`	Small-molecule structure	小分子结构	—	Small-molecule graphs, SMILES strings, fingerprints, or molecular structure records.
`structure-3d`	3D structure	三维结构	—	Molecular coordinates
`raw-omics`	Raw omics	原始组学数据	—	Raw or minimally processed omics data.
`image`	Image	图像	—	General scientific image input.
`database`	Database	数据库	—	Structured external scientific databases.
`web`	Web	网页	—	Browser or URL-provided evidence.
`code`	Code	代码	—	Source code
`wet-lab-output`	Wet-lab output	湿实验输出	—	Instrument output or laboratory measurements.

Access level

ID	English	中文	Parent	Definition
`fully-open`	Fully open	完全公开	—	Tasks and required evaluation materials are publicly accessible.
`partially-open`	Partially open	部分公开	—	Some tasks
`metadata-only`	Metadata only	仅元数据	—	Public descriptions exist but runnable evaluation materials are unavailable.
`private-internal`	Private or internal	私有或内部	—	The benchmark is reported but not publicly released.

Scientific object

ID	English	中文	Parent	Definition
`protein`	Protein	蛋白质	—	Protein sequences
`dna`	DNA	DNA	—	DNA sequences
`rna`	RNA	RNA	—	RNA sequences
`small-molecule`	Small molecule	小分子	—	Ligands
`cell`	Cell	细胞	—	Individual cells
`tissue`	Tissue	组织	—	Tissues
`organism-population`	Organism or population	个体或群体	—	Organisms
`microbial-community`	Microbial community	微生物群落	—	Microbial communities
`omics-profile`	Omics profile	组学谱	—	Genome-scale molecular measurement matrices and integrated omics profiles.
`experimental-system`	Experimental system	实验系统	—	Assays

Scientific task family

ID	English	中文	Parent	Definition
`structure-prediction`	Structure prediction	结构预测	—	Infer a molecular structure
`design-generation`	Design and generation	设计与生成	—	Generate or optimize a molecular sequence
`molecular-interaction`	Molecular interaction	分子相互作用	—	Predict interaction
`property-function-prediction`	Property and function prediction	性质与功能预测	—	Predict molecular properties
`sequence-regulation`	Sequence and regulation	序列与调控	—	Identify functional sequence elements
`variant-perturbation`	Variant and perturbation effect	变异与扰动效应	—	Predict the consequence of a mutation
`omics-analysis`	Omics analysis	组学分析	—	Analyze molecular profiles across samples
`cellular-spatial-analysis`	Cellular and spatial analysis	细胞与空间分析	—	Analyze cell identity
`statistical-genetics`	Statistical genetics	统计遗传学	—	Analyze association
`scientific-workflow`	Scientific workflow	科学工作流	—	Retrieve
`systems-modeling`	Systems modeling	系统建模	—	Reconstruct

Scientific task

ID	English	中文	Parent	Definition
`protein-structure`	Protein structure prediction	蛋白质结构预测	—	Predict or assess the three-dimensional structure of proteins or protein assemblies.
`protein-monomer-structure-prediction`	Protein monomer structure prediction	蛋白质单体折叠与结构预测	protein-structure	Predict the three-dimensional structure of a single protein chain or domain.
`protein-complex-structure-prediction`	Protein complex structure prediction	蛋白质复合物结构预测	protein-structure	Predict the assembled structure or stoichiometry of a protein complex.
`protein-model-quality-assessment`	Protein model quality assessment	蛋白质模型质量评估	protein-structure	Estimate the accuracy or quality of a predicted protein structure model.
`protein-secondary-structure-prediction`	Protein secondary-structure prediction	蛋白质二级结构预测	protein-structure	Predict per-residue helix
`protein-contact-map-prediction`	Protein contact-map prediction	蛋白质接触图预测	protein-structure	Predict which residue pairs in a protein are spatially in contact.
`protein-design`	Protein design	蛋白质设计	—	Generate or optimize proteins under structural
`protein-sequence-design`	Protein sequence design	蛋白质序列设计	protein-design	Generate an amino-acid sequence satisfying specified structural or functional constraints.
`protein-fitness-optimization`	Protein fitness optimization	蛋白质适应度优化	protein-design	Select or optimize protein variants toward improved measured fitness.
`protein-property-function`	Protein property and function prediction	蛋白质性质与功能预测	—	Predict protein function
`protein-function-annotation`	Protein function annotation	蛋白质功能注释	protein-property-function	Assign molecular function
`protein-remote-homology-detection`	Protein remote-homology detection	蛋白质远缘同源检测	protein-property-function	Detect structural or evolutionary relationships between distantly related protein sequences.
`protein-residue-identity-prediction`	Protein residue-identity prediction	蛋白质残基身份预测	protein-property-function	Predict the amino-acid identity associated with a local protein structural environment.
`protein-fluorescence-prediction`	Protein fluorescence prediction	蛋白质荧光性质预测	protein-property-function	Predict a measured fluorescence phenotype from a protein sequence.
`protein-mutation-effect-prediction`	Protein mutation-effect prediction	蛋白质突变效应预测	protein-property-function	Predict the functional or phenotypic effect of protein sequence variants.
`protein-fitness-prediction`	Protein fitness prediction	蛋白质适应度预测	protein-property-function	Predict a quantitative or ranked protein fitness measurement from sequence.
`protein-stability-prediction`	Protein stability prediction	蛋白质稳定性预测	protein-property-function	Predict stability
`protein-solubility-prediction`	Protein solubility prediction	蛋白质溶解性预测	protein-property-function	Predict whether or how strongly a protein is soluble under an assay condition.
`protein-clinical-variant-interpretation`	Protein clinical variant interpretation	蛋白质临床变异解读	protein-property-function	Classify or prioritize protein variants for clinical relevance or pathogenicity.
`molecular-interaction-analysis`	Molecular interaction and binding	分子相互作用与结合	—	Predict or analyze interactions
`protein-protein-interaction-prediction`	Protein-protein interaction prediction	蛋白质-蛋白质相互作用预测	molecular-interaction-analysis	Predict whether proteins interact or identify an interaction partner or interface.
`protein-protein-interface-prediction`	Protein-protein interface prediction	蛋白质-蛋白质界面预测	molecular-interaction-analysis	Predict which residues from two proteins will contact one another when the proteins bind.
`protein-complex-mutation-stability-prediction`	Protein-complex mutation stability prediction	蛋白质复合物突变稳定性预测	molecular-interaction-analysis	Predict whether a mutation increases or decreases the stability of a protein complex or interaction.
`protein-protein-binding-affinity`	Protein-protein binding affinity	蛋白质-蛋白质结合亲和力	molecular-interaction-analysis	Predict or rank the strength of binding between protein partners.
`antibody-antigen-interaction`	Antibody-antigen interaction	抗体-抗原相互作用	molecular-interaction-analysis	Predict antibody-antigen recognition
`protein-ligand-binding-prediction`	Protein-ligand binding prediction	蛋白质-配体结合预测	molecular-interaction-analysis	Predict or reason about protein-ligand binding when the official source does not distinguish pose from affinity.
`protein-ligand-efficacy-prediction`	Protein-ligand functional efficacy prediction	蛋白质-配体功能效应预测	molecular-interaction-analysis	Given a ligand and target structures, predict whether the ligand activates or inhibits the protein's function rather than whether it binds.
`protein-ligand-pose-prediction`	Protein-ligand pose prediction	蛋白质-配体结合构象预测	molecular-interaction-analysis	Predict the bound pose or geometry of a small molecule in a protein complex.
`protein-ligand-binding-affinity`	Protein-ligand binding affinity	蛋白质-配体结合亲和力	molecular-interaction-analysis	Predict or rank the strength of protein-small-molecule binding.
`rna-protein-interaction-prediction`	RNA-protein interaction prediction	RNA-蛋白质相互作用预测	molecular-interaction-analysis	Predict whether an RNA and protein interact or bind.
`dna-regulation-perturbation`	DNA regulation and perturbation	DNA调控与扰动	—	Predict DNA regulatory elements
`promoter-detection`	Promoter detection	启动子识别	dna-regulation-perturbation	Identify promoter regions or promoter activity from DNA sequence.
`enhancer-activity-prediction`	Enhancer activity prediction	增强子活性预测	dna-regulation-perturbation	Predict enhancer activity or regulatory potential from DNA sequence.
`transcription-factor-binding-site-prediction`	Transcription-factor binding-site prediction	转录因子结合位点预测	dna-regulation-perturbation	Predict transcription-factor binding sites or occupancy from sequence or genomic context.
`enhancer-promoter-interaction-prediction`	Enhancer-promoter interaction prediction	增强子-启动子相互作用预测	dna-regulation-perturbation	Predict physical or functional enhancer-promoter interactions.
`epigenetic-mark-prediction`	Epigenetic-mark prediction	表观遗传标记预测	dna-regulation-perturbation	Predict DNA methylation
`dna-variant-effect-prediction`	DNA variant-effect prediction	DNA变异效应预测	dna-regulation-perturbation	Predict regulatory
`dna-sequence-analysis`	DNA sequence analysis	DNA序列分析	dna-regulation-perturbation	Analyze DNA sequence properties
`crispr-guide-activity-prediction`	CRISPR guide activity prediction	CRISPR向导活性预测	dna-regulation-perturbation	Predict on-target activity or efficiency of a CRISPR guide sequence.
`crispr-off-target-prediction`	CRISPR off-target prediction	CRISPR脱靶预测	dna-regulation-perturbation	Predict unintended CRISPR guide interactions or editing activity at off-target sequences.
`dna-sequence-design`	DNA sequence design	DNA序列设计	dna-regulation-perturbation	Generate or optimize a DNA sequence for a specified regulatory or experimental function.
`rna-function-design`	RNA function and design	RNA功能与设计	—	Predict or design RNA structure
`rna-structure-prediction`	RNA structure prediction	RNA结构预测	rna-function-design	Predict the secondary or tertiary structure of an RNA molecule.
`rna-structure-quality-assessment`	RNA structure quality assessment	RNA结构质量评估	rna-function-design	Score or rank candidate RNA tertiary structures by their expected similarity to the native structure.
`rna-function-classification`	RNA function classification	RNA功能分类	rna-function-design	Assign functional or biotype labels to an RNA sequence.
`rna-modification-prediction`	RNA modification prediction	RNA修饰预测	rna-function-design	Predict RNA modification types or sites from sequence or context.
`rna-processing-translation-prediction`	RNA processing and translation prediction	RNA加工与翻译预测	rna-function-design	Predict RNA processing
`rna-stability-degradation-prediction`	RNA stability and degradation prediction	RNA稳定性与降解预测	rna-function-design	Predict RNA stability
`rna-switch-activity-prediction`	Programmable RNA-switch activity prediction	可编程RNA开关活性预测	rna-function-design	Predict functional activity or expression measurements of engineered RNA switches.
`rna-design`	RNA design	RNA设计	rna-function-design	Generate or optimize an RNA sequence for a specified functional response.
`sirna-efficacy-prediction`	siRNA efficacy prediction	siRNA效能预测	rna-function-design	Predict the silencing efficacy of an siRNA sequence or sequence pair.
`small-molecule-discovery`	Small-molecule discovery	小分子发现	—	Generate
`small-molecule-generation`	Small-molecule generation	小分子生成	small-molecule-discovery	Generate molecular structures satisfying requested constraints or objectives.
`small-molecule-property-prediction`	Small-molecule property prediction	小分子性质预测	small-molecule-discovery	Predict physicochemical or biological properties of a small molecule.
`admet-toxicity-prediction`	ADMET and toxicity prediction	ADMET与毒性预测	small-molecule-discovery	Predict absorption
`reaction-prediction`	Chemical reaction prediction	化学反应预测	small-molecule-discovery	Predict products
`retrosynthesis-planning`	Retrosynthesis planning	逆合成规划	small-molecule-discovery	Propose precursor and reaction routes for synthesizing a target molecule.
`omics-cellular-analysis`	Omics and cellular analysis	组学与细胞分析	—	Analyze genome-scale profiles
`differential-expression-analysis`	Differential expression analysis	差异表达分析	omics-cellular-analysis	Identify genes or features with condition-associated expression changes.
`differential-abundance-analysis`	Differential abundance analysis	差异丰度分析	omics-cellular-analysis	Identify features
`cell-type-annotation`	Cell-type annotation	细胞类型注释	omics-cellular-analysis	Assign biological cell-type or state labels to single-cell profiles.
`cell-state-clustering`	Cell-state clustering	细胞状态聚类	omics-cellular-analysis	Discover cell populations or states by clustering molecular profiles.
`single-cell-data-integration`	Single-cell data integration	单细胞数据整合	omics-cellular-analysis	Integrate single-cell datasets across batches
`trajectory-inference`	Trajectory inference	轨迹推断	omics-cellular-analysis	Infer developmental
`gene-regulatory-network-inference`	Gene-regulatory-network inference	基因调控网络推断	omics-cellular-analysis	Infer transcription-factor and target-gene regulatory relationships.
`pathway-enrichment-analysis`	Pathway enrichment analysis	通路富集分析	omics-cellular-analysis	Identify pathways or gene sets enriched in a list or ranked molecular profile.
`spatial-omics-analysis`	Spatial omics analysis	空间组学分析	omics-cellular-analysis	Analyze spatially resolved molecular profiles
`multiomics-integration`	Multi-omics integration	多组学整合	omics-cellular-analysis	Integrate two or more molecular modalities into a joint analysis or representation.
`biomarker-discovery`	Biomarker discovery	生物标志物发现	omics-cellular-analysis	Identify molecular features or signatures associated with phenotype
`microbial-metagenomic-analysis`	Microbial and metagenomic analysis	微生物与宏基因组分析	omics-cellular-analysis	Analyze microbial composition
`statistical-population-genetics-analysis`	Statistical and population genetics	统计与群体遗传分析	—	Analyze genetic association
`genetic-association-causal-mapping`	Genetic association and causal mapping	遗传关联与因果定位	statistical-population-genetics-analysis	Identify genetic associations
`heritability-polygenic-prediction`	Heritability and polygenic prediction	遗传力与多基因预测	statistical-population-genetics-analysis	Estimate trait architecture
`population-genetics-analysis`	Population genetics analysis	群体遗传分析	statistical-population-genetics-analysis	Analyze ancestry
`scientific-workflow-systems`	Scientific workflow and systems analysis	科学工作流与系统分析	—	Retrieve evidence
`scientific-database-retrieval`	Scientific database retrieval	科学数据库检索	scientific-workflow-systems	Retrieve a scientific record
`scientific-evidence-interpretation`	Scientific evidence interpretation	科学证据解读	scientific-workflow-systems	Interpret or reconcile scientific text
`experiment-protocol-planning`	Experiment and protocol planning	实验与方案规划	scientific-workflow-systems	Design an experiment
`end-to-end-computational-analysis`	End-to-end computational analysis	端到端计算分析	scientific-workflow-systems	Complete a multi-step scientific analysis using data
`reaction-network-reconstruction`	Reaction-network reconstruction	反应网络重建	scientific-workflow-systems	Reconstruct a biochemical reaction network or executable systems model.
`simulation-based-experiment`	Simulation-based experiment	仿真实验	scientific-workflow-systems	Interact with or reason over a simulator to identify