A pipeline for annotation of genes with DeepFRI, a deep learning model for functional protein annotation with Gene Ontology (GO) terms. It incorporates FoldComp databases of predicted protein structures for fast annotation of metagenomic gene catalogues.
Proteins perform most of the work of living cells. Amino acid sequence and structural features of proteins determine a wide range of functions: from binding specificity and conferring mechanical stability, to catalysis of biochemical reactions, transport, and signal transduction. DeepFRI is a neural network designed to predict protein function within the framework of the Gene Ontology (GO). The exponential growth in the number of available protein sequences, driven by advancements in low-cost sequencing technologies and computational methods (e.g. gene prediction), has resulted in a pressing need for efficient software to facilitate the annotation of protein databases. Metagenomic-DeepFRI addresses such needs, building upon efficient libraries. It incorporates novel databases of predicted structures (AlphaFold, ESMFold, MIP, etc.) and improves runtimes of DeepFRI by 2-12 times!
- Search proteins similar to query in PDB and supply
FoldCompdatabases withMMSeqs2. - Find the best alignment among
MMSeqs2hits usingPyOpal. - Align target protein contact map to query protein with unknown structure.
- Run
DeepFRIwith the structure if found in the database, otherwise runDeepFRIwith sequence only.
- Install from PyPI. Installation might take a few minutes due to download of MMseqs2 binaries.
pip install mdeepfri
- Run and view the help message.
mDeepFRI --help
The PDB database will be automatically downloaded and installed during the first run of mDeepFRI. The PDB suffers from formatting inconsistencies, therefore during PDB alignment around 10% will fail and will be reported via WARNING. We suggest coupling PDB search with predicted databases, as it massively improves the structural coverage of the protein universe. A good protein structure allows DeepFRI to annotate the function in more detail. However, the sequence branch of the model has the largest weight, thus even if the predicted structure is erroneous, it will have a minor effect on the prediction. The details can be found in the original manuscript, fig. 2A.
You can download additional databases from website. During a first run, FASTA sequences will be extracted from FoldComp database and MMseqs2 database will be created and indexed. You can use different databases, but be mindful that computation time might increase exponentially with the size of the database.
Tested databases:
afdb_swissprotafdb_swissprot_v4afdb_rep_v4afdb_rep_dark_v4afdb_uniprot_v4esmatlasesmatlas_v2023_02highquality_clust30
ATTENTION: Please, do not rename downloaded databases. FoldComp has certain inconsistencies in the way FASTA sequences are extracted (example), therefore pipeline was tweaked for each database. If database you need does not work, please report in issues and we will add it as soon as possible. Sorry for the inconvenience.
ATTENTION: database creation is a very sensitive step which relies on external software. If pipeline is interrupted during this step, the databases might be corrupted. If you are not sure about your database, rerun the pipeline with --overwrite flag - it will rerun database creation process.
In order to use personal database of structures, you will have to create a custom FoldComp database. For that, download a FoldComp executable and run the following command:
foldcomp compress [-t number] <dir|tar(.gz)> [<dir|tar|db>]
Two versions of models available:
v1.0- is the original version from DeepFRI publication.v1.1- is a version finetuned on AlphaFold models and machine-generated Gene Ontology Uniprot annotations. You can read details aboutv1.1in ISMB 2023 presentation by Pawel Szczerbiak
To download models run command:
mDeepFRI get-models --output path/to/weights/folder -v {1.0 or 1.1}
mDeepFRI predict-function -i /path/to/protein/sequences -d /path/to/foldcomp/database/ -w /path/to/deepfri/weights/folder -o /output_path 2> log.txt
The logging module writes output into stderr, so use 2> to redirect it to the file.
Other available parameters can be found upon command mDeepFRI --help.
The output folder will contain:
{database_name}.search_results.tsvquery.mmseqsDB+ index from MMSeqs2 search.results.tsv- a final output from the DeepFRI model.
| Protein | GO_term/EC_numer | Score | Annotation | Neural_net | DeepFRI_mode | DB_hit | DB_name | Identity |
|---|---|---|---|---|---|---|---|---|
| MIP_00215364 | GO:0016798 | 0.218 | hydrolase activity, acting on glycosyl bonds | gcn | mf | MIP_00215364 | mip_rosetta_hq | 0.933 |
| 1GVH_1 | GO:0009055 | 0.217 | electron transfer activity | gnn | mf | AF-P24232-F1-model_v4 | afdb_swissprot_v4 | 1.0 |
| unaligned | 3.2.1.- | 0.215 | 3.2.1.- | cnn | ec | nan | nan | nan |
This is an example of protein annotation with the AlphaFold database.
- Protein - the name of the protein from the FASTA file.
- GO_term/EC_numer - predicted GO term or EC number (dependent on mode)
- Score - DeepFRI score, translates to model confidence in prediction. Details in publication.
- Annotation - annotation from ontology
- Neural_net - type of neural network used for prediction (gcn = Graph Convolutional Network; cnn = Convolutional Neural Network). GCN (Graph Convolutional Network) is used when structural information is available in the database, allowing for generally more confident predictions. When there are no proteins above similarity cut-off (50% identity by default), CNN is used.
- DeepFRI_mode:
mf = molecular_function bp = biological_process cc = cellular_component ec = enzyme_commission - DB_hit - name of the hit in the database. Empty if no hit was found.
- DB_name - name of the database. Empty if no hit was found.
- Identity - sequence identity between query and hit. Empty if no hit was found.
The GO ontology contains three subontologies, defined by their root nodes:
- Molecular Function (MF)
- Biological Process (BP)
- Cellular Component (CC)
- Additionally, Metagenomic-DeepFRI v1.0 is able to predict Enzyme Comission number (EC).
By default, the tool makes predictions in all 4 categories. To select only a few pass the parameter
-por--processing-modesfew times, i.e.:
mDeepFRI predict-function -i /path/to/protein/sequences -d /path/to/foldcomp/database/ -w /path/to/deepfri/weights/folder -o /output_path -p mf -p bp
Different databases have a different level of evidence. For example, PDB structures are real experimental structures, thus they are considered to be the data of highest quality. Therefore new proteins are first queried against PDB. Computational predictions differ by quality, i.e. AlphaFold predictions are often more accurate than ESMFold predictions. We provide an opporunity to search multiple databases in a hierarchical manner. For example, if you want to search AlphaFold database first, and then ESMFold, you can pass the parameter -d or --databases few times, i.e.:
mDeepFRI predict-function -i /path/to/protein/sequences -d /path/to/alphafold/database/ -d /path/to/another/esmcomp/database/ -w /path/to/deepfri/weights/folder -o /output_path
The first run of mDeepFRI with the database will create temporary files, needed for the pipeline. If you don't want to keep them for the next run add
flag --remove-intermediate.
If argument threads is provided, the app will parallelize certain steps (alignment, contact map alignment, functional annotation).
GPU is often used to speed up neural networks. Metagenomic-DeepFRI takes care of this and, if CUDA is installed on your machine, mDeepFRI will automatically use it for prediction. If not, the model will use CPUs.
Technical tip: Single instance of DeepFRI on GPU requires 2GB VRAM. Every currently available GPU with CUDA support should be able to run the model.
Metagenomic-DeepFRI is a scientific software. If you use it in an academic work, please cite the papers behind it:
- Gligorijević et al. "Structure-based protein function prediction using graph convolutional networks" Nat. Comms. (2021). https://doi.org/10.1038/s41467-021-23303-9
- Steinegger & Söding "MMseqs2 enables sensitive protein sequence searching for the analysis of massive data sets" Nat. Biotechnol. (2017) https://doi.org/10.1038/nbt.3988
- Kim, Midrita & Steinegger "Foldcomp: a library and format for compressing and indexing large protein structure sets" Bioinformatics (2023) https://doi.org/10.1093/bioinformatics/btad153
- Maranga et al. "Comprehensive Functional Annotation of Metagenomes and Microbial Genomes Using a Deep Learning-Based Method" mSystems (2023) https://doi.org/10.1128/msystems.01178-22
Found a bug ? Have an enhancement request ? Head over to the GitHub issue tracker if you need to report or ask something. If you are filing in on a bug, please include as much information as you can about the issue, and try to recreate the same bug in a simple, easily reproducible situation.
Contributions are more than welcome! See
CONTRIBUTING.md
for more details.
This project adheres to Semantic Versioning and provides a changelog in the Keep a Changelog format.
This library is provided under the The 3-Clause BSD License.