Biases in the Experimental Annotations of Protein Function and Their Effect on Our Understanding of Protein Function Space
Open accessThe DOI matches its Crossref record; 53 of the 53 references it cites were found in a registry or an index.
Abstract
The ongoing functional annotation of proteins relies upon the work of curators to capture experimental findings from scientific literature and apply them to protein sequence and structure data. However, with the increasing use of high-throughput experimental assays, a small number of experimental studies dominate the functional protein annotations collected in databases. Here, we investigate just how prevalent is the "few articles - many proteins" phenomenon. We examine the experimentally validated annotation of proteins provided by several groups in the GO Consortium, and show that the distribution of proteins per published study is exponential, with 0.14% of articles providing the source of annotations for 25% of the proteins in the UniProt-GOA compilation. Since each of the dominant articles describes the use of an assay that can find only one function or a small group of functions, this leads to substantial biases in what we know about the function of many proteins. Mass-spectrometry, microscopy and RNAi experiments dominate high throughput experiments. Consequently, the functional information derived from these experiments is mostly of the subcellular location of proteins, and of the participation of proteins in embryonic developmental pathways. For some organisms, the information provided by different studies overlap by a large amount. We also show that the information provided by high throughput experiments is less specific than those provided by low throughput experiments. Given the experimental techniques available, certain biases in protein function annotation due to high-throughput experiments are unavoidable. Knowing that these biases exist and understanding their characteristics and extent is important for database curators, developers of function annotation programs, and anyone who uses protein function annotation data to plan experiments.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures




Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Moreover, increasing numbers of functional annotations are based on phenotypes from high-throughput screens for genetic phenotypes or protein–protein interactions, both of which are prone to generating false positives [100].”
From Functional unknomics: Systematic screening of conserved genes of unknown function (2023) · cites this paper as [100]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Gapped BLAST and PSI-BLAST: a new generation of protein database search programs
Altschul, S. (1997) · Nucleic Acids Research · doi:10.1093/nar/25.17.3389 · Cited by 73,302 · cited 5×
“Count of ECO terms in top-50 papers N ECO term ECO ID Articles 1 protein separation followed by fragment iden- tification evidence ECO:0000160 27 2 sequence similarity evidence ECO:0000044 27 3 imaging assay evidence ECO:0000324 24 4 cell fractionation evidence ECO:0000004 23 5 transmembrane domain prediction ev- iden…”
InfluentialGene Ontology: tool for the unification of biology
Ashburner, M., Ball, C. A., Blake, J. A., Botstein, D., Butler, H., Cherry, J. M., Davis, A. P., Dolinski, K., Dwight, S. S., Eppig, J. T., et al. (2000) · Nature Genetics · doi:10.1038/75556 · Cited by 43,206
Matplotlib: a 2D graphics environment
Hunter, J. D. (2007) · Computing in Science & Engineering · doi:10.1109/mcse.2007.55 · Cited by 39,249 · cited 1×
“The following tools were used in the analyses: Biopython [20], matplotlib [21].”
Biopython: freely available python tools for computational molecular biology and bioinformatics
Cock, P. J. A., Antao, T., Chang, J. T., Chapman, B. A., Cox, C. J., Dalke, A., Friedberg, I., Hamelryck, T., Kauff, F., Wilczynski, B., et al. (2009) · Bioinformatics · doi:10.1093/bioinformatics/btp163 · Cited by 6,080 · cited 3×
“The following tools were used in the analyses: Biopython [20], mat349 plotlib [21].”
InfluentialSelf-citationGlobal analysis of protein localization in budding yeast
Huh, W.-K., Falvo, J. V., Gerke, L. C., Carroll, A. S., Howson, R. W., Weissman, J. S., & O'Shea, E. K. (2003) · Nature · doi:10.1038/nature02026 · Cited by 4,248
Systematic functional analysis of the Caenorhabditis elegans genome using RNAi
Kamath, R. S., Fraser, A. G., Dong, Y., Poulin, G., Durbin, R., Gotta, M., Kanapin, A., Le Bot, N., Moreno, S., Sohrmann, M., et al. (2003) · Nature · doi:10.1038/nature01278 · Cited by 3,616
Towards a proteome-scale map of the human protein–protein interaction network
Rual, J.-F., Venkatesan, K., Hao, T., Hirozane-Kishikawa, T., Dricot, A., Li, N., Berriz, G. F., Gibbons, F. D., Dreze, M., Ayivi-Guedehoussou, N., et al. (2005) · Nature · doi:10.1038/nature04209 · Cited by 2,933
Proteome survey reveals modularity of the yeast cell machinery
Gavin, A.-C., Aloy, P., Grandi, P., Krause, R., Boesche, M., Marzioch, M., Rau, C., Jensen, L. J., Bastuck, S., Dümpelfeld, B., et al. (2006) · Nature · doi:10.1038/nature04532 · Cited by 2,575
Genes required for mycobacterial growth defined by high density mutagenesis
Sassetti, C. M., Boyd, D. H., & Rubin, E. J. (2003) · Molecular Microbiology · doi:10.1046/j.1365-2958.2003.03425.x · Cited by 2,544
A Mitochondrial Protein Compendium Elucidates Complex I Disease Biology
Pagliarini, D. J., Calvo, S. E., Chang, B., Sheth, S. A., Vafai, S. B., Ong, S.-E., Walford, G. A., Sugiana, C., Boneh, A., Chen, W. K., et al. (2008) · Cell · doi:10.1016/j.cell.2008.06.016 · Cited by 1,994
Cited by
Items 101–134 of 134
Organizing the bacterial annotation space with amino acid sequence embeddings
Grigson, S., McKerral, J. C., Mitchell, J. G., & Edwards, R. (2022) · BMC Bioinformatics · doi:10.1186/s12859-022-04930-5 · Cited by 3
“Additionally, annotations from high-throughput experiments are often strongly biased towards a limited number of functions, leaving large regions of the protein space unexplored [12].”
BackgroundRoche, D. B., & Brüls, T. (2015) · Protein Science · doi:10.1002/pro.2635 · Cited by 4
“26 Even though some concerns have been expressed over the overall accuracy of some annotation protocols, 3,27 using the Uniprot—Swissprot and TrEMBL—resources for EC labelling is probably a reasonable choice given the vast majority of proteins arising from sequencing projects do not have any…”
MethodEvoWeaver: large-scale prediction of gene functional associations from coevolutionary signals
Lakshman, A., & Wright, E. S. (2025) · Nature Communications · doi:10.1038/s41467-025-59175-6 · Cited by 4
“The sparsity of high-quality annotations exacerbates the problem of non-speci fi c and low-con fi dence annotations that proliferate across genomes 8,9 .”
Molecular biotechnology for the biodegradation of organofluorine compounds
Wackett, L. P. (2026) · Current Opinion in Biotechnology · doi:10.1016/j.copbio.2026.103446 · Cited by 3
de Siqueira, G. M. V., Eng, T., Mukhopadhyay, A., & Guazzaroni, M. (2025) · mSphere · doi:10.1128/msphere.00391-25 · Cited by 3
“…and given that annotation quality often relies on a small subset of experiments which cannot capture the full extent of protein functions (39, 40), we believe that these results might inform neglected aspects of the P. putida KT2440 genome that could benefit from efforts to provide further…”
CELLE-2: Translating Proteins to Pictures and Back with a Bidirectional Text-to-Image Transformer
Khwaja, E., Song, Y., Agarunov, A., & Huang, B. (2023) · neural information processing systems · doi:10.52202/075280-0217 · Cited by 5
A continuous assay for l-talarate/galactarate dehydratase using circular dichroism
Easton, N. M., Aboushawareb, S. A. E., & Bearne, S. L. (2017) · Analytical Biochemistry · doi:10.1016/j.ab.2017.12.015 · Cited by 3
Phan, A., Joshi, P., Kadelka, C., & Friedberg, I. (2025) · Database · doi:10.1093/database/baaf036 · Cited by 2
“As depicted in Figure 1A, articles describing high-throughput experiments consistently yield less informative annotations for the proteins, which aligns with previous findings (50, 55).”
Self-citationLu, M., & Liu, T. (2025) · Analytical Biochemistry · doi:10.1016/j.ab.2025.116005 · Cited by 2
MSF: Modulated Sub-graph Finder
Farman, M. R., Hofacker, I. L., & Amman, F. (2019) · F1000Research · doi:10.12688/f1000research.16005.3 · Cited by 2
Assunção, C. B., de Aguiar, E. L., Chávez‐Fumagalli, M. Á., Santrer, E. F. R. L., Dias, S. R., Rodrigues, T. D. S., & Caligiorne, R. B. (2024) · Current Biotechnology · doi:10.2174/0122115501285434240409040348 · Cited by 2
Identifying noisy functional annotations of proteins using sparse semantic similarity
Zhiwen, Y. U., Xia, C. H. E. N., Wang, J., Yu, G., & Chang, L. U. (2017) · Scientia Sinica Informationis · doi:10.1360/n112017-00105 · Cited by 2
Scapolatiello, A., Manfrin, C., Greco, S., Rončević, T., Pallavicini, A., Puljas, S., & Gerdol, M. (2023) · Diversity · doi:10.3390/d15060707 · Cited by 2
“Such dynamics, summarized in Figure 3, have been previously well documented by other authors and are key determinants for the progression of annual life cycles in C. kusceri , influencing both reproduction and growth [11,12,20].”
ResultA kaleidoscopic look at vanillyl alcohol oxidase
Gygli, G. (2018) · doi:10.18174/431725 · Cited by 2
Machine Learning for Sequence‐to‐Function Approaches
Barghout, R. A., Kirby, M., Zheng, A., Chinas, L., Mohammadi, M., Xu, Z., Sánchez-Lengeling, B., & Mahadevan, R. (2026) · Wiley eBooks · doi:10.1002/9783527850532.ch4 · Cited by 2
Yin, W., Wang, Q., Zeng, Z., & Xie, X. (2026) · Frontiers in Molecular Biosciences · doi:10.3389/fmolb.2026.1818173 · Cited by 1
“This “few studies, many annotations” phenomenon affects peptide resources as well and promotes overfitting of models to well studied sequence families (Schnoes et al., 2013 Second, construction of negative samples for supervised learning carries major pitfalls.”
MSF: Modulated Sub-graph Finder
Farman, M. R., Hofacker, I. L., & Amman, F. (2018) · F1000Research · doi:10.12688/f1000research.16005.1 · Cited by 1
Srivastava, H., Berenberg, D., Qassab, O., Wang, Z., Bonneau, R., & Carlton, J. M. (2025) · bioRxiv · doi:10.1101/2025.09.12.675843 · Cited by 1
Pharmacolgical and biological annotations enhance functional residues prediction
Maietta, P. (2017) · LA Referencia (Red Federada de Repositorios Institucionales de Publicaciones Científicas)
Data-driven prediction of molecular function
Litfin, T. (2020) · Griffith Research Online · doi:10.25904/1912/29
Tao, J., Brayton, K. A., & Broschat, S. L. (2021) · Frontiers in Bioinformatics · doi:10.3389/fbinf.2021.749008
“Unfortunately, these methods are time-consuming and require the use of substantial resources due to their low throughput and restricted scope of methodology (Schnoes et al., 2013; Cao and Cheng, 2016; Frasca and Cesa-Bianchi, 2017; Hong et al., 2020).”
BackgroundΓαλλιοπούλου, Ε. (2018) · doi:10.12681/eadd/43759 · Cited by 1
Jia, B., Feng, T., Li, C., Huang, B., Hao, F., & Liu, D. (2023) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm58861.2023.10385542
Additional file 1 of Organizing the bacterial annotation space with amino acid sequence embeddings
Grigson, S. R., McKerral, J. C., Mitchell, J. G., & Edwards, R. A. (2022) · Figshare · doi:10.6084/m9.figshare.21200740
Wira, J., Judd, M., Baldino, A., Bachvaroff, T. R., & Place, A. R. (2026) · Journal of Phycology · doi:10.1111/jpy.70152 · Cited by 1
“…biochemical assays, and expression ex-periments—that can both clarify protein roles and help expand the databases on which computational methods rely and may also provide novel insights and mechanisms for optogenetics (Awasthi et al., 2020; Roberts et al., 2011; Schnoes et al., 2013).”
CB-Search: a method for searching for similar protein motifs with biased compositions
Jarnot, P. (2026) · BMC Bioinformatics · doi:10.1186/s12859-026-06509-w
“Nevertheless, when comparing search tools, all of them are equally affected by this noise [46].”
AUKAT: Conditional VAE-Driven Augmentation and Neural Modeling of Enzyme Turnover Numbers
Liu, M., Ni, X., & Bryliński, M. (2026) · Biomolecules · doi:10.3390/biom16071049 · Cited by 2
Pho, V.-S., Bianchi, A. N., Scarsini, M., Bowler, C., & Carbone, A. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.07.31.742035
Interpretable prediction of nucleic acid-binding proteins using a protein language model
Kim, H., Lee, S.-G., Oh, J.-S., Kim, K. K., Kim, E.-M., & Park, C. (2026) · Bioinformatics Advances · doi:10.1093/bioadv/vbag235
“These sources can introduce noise through in silico annotation and lead to circular validation, in which models learn to reproduce biases from earlier prediction tools rather than true biological patterns (Schnoes et al. 2013).”
BIOLOGICAL PROCESS ANNOTATION OF PROTEINS ACROSS THE KINGDOM PLANTAE
ODE P.O., K. (2020) · International Journal of Multidisciplinary Research and Studies
In Other Words: Learning When to Trust Generated Protein Function DescriptionsPreprint
Li, T., Ram, P., Tawfiq, R., Guintu, F., Yang, C., & Bromberg, Y. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.10.15.757758
“…can preserve sequence and/or structure while shifting functional scope, regulation, or context of action [17], (2) convergent evolution can produce similar functions with minimal similarity [18], and (3) existing experimental biases limit annotations to a narrow slice of the protein space [19].”
FAIRsharing record for: Enzyme Structure Function Ontology
FAIRsharing Team (2018) · FAIRsharing.org · doi:10.25504/fairsharing.9y3gv0
The chemistry and evolution of enzyme function: isomerases as a case study
Cuesta, S. M. (2014) · Apollo (University of Cambridge) · doi:10.17863/cam.15989
Metadata
- Authors
- Schnoes, A. M., Ream, D. C., Thorman, A. W., Babbitt, P. C., & Friedberg, I.
- Affiliations
- University of California, San Francisco (US)
- Miami University (US)
- Corresponding author
- Friedberg, Iddo
- Venue
- PLoS Computational Biology
- Publisher
- Public Library of Science (PLoS)
- Published
- 2013-05-30
- Volume
- 9
- Issue
- 5
- Pages
- e1003063
- ISSN
- 1553-734X (ISSN-L)
- 1553-7358
- PMC
- PMC3667760
- OpenAlex
- W2169059428
- Cited by
- 129
- References
- 53
- Citations by year
- 2026: 9 · 2025: 10 · 2024: 3 · 2023: 12 · 2022: 9 · 2021: 7 · 2020: 7 · 2019: 12 · 2018: 9 · 2017: 16 · 2016: 12 · 2015: 11 · 2014: 11 · 2013: 1
- FWCI
- 5.23
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- PLoS Computational Biology (publishedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- eScholarship (California Digital Library) (submittedVersion)
- eScholarship (California Digital Library) (submittedVersion, cc-by)
- arXiv (Cornell University)
- Funders
- National Science Foundation (1146960, 0640476)
- National Institutes of Health (NIH-R01)
- National Institute of General Medical Sciences (R01GM60595, R01GM060595)
- Division of Biological Infrastructure
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Bioinformatics and Genomic Networks
- Health Sciences › Medicine › Health Informatics › Biomedical Text Mining and Ontologies
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
- Keywords
- protein annotation, annotation bias, high-throughput experimentation, Gene Ontology, subcellular localization, embryonic development, mass spectrometry, RNA interference, literature curation
- MeSH terms
- Animals, Humans, Proteins, Computational Biology, Databases, Protein, High-Throughput Screening Assays, Molecular Sequence Annotation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.