Rick Masonbrink

Rick Masonbrink

I am fortunate to work as a research scientist in the Genome Informatics Facility at Iowa State University, where I am surrounded by highly skilled bioinformaticians and constantly exposed to new ideas and approaches. This dynamic environment fosters my passion for science and my curiosity to keep learning and growing. With the support of the USDA, I have had the opportunity to create workshops and tutorials, and I also assist our faculty and industry collaborators through fee-for-service work. Over the years, I’ve been involved in a wide range of projects, developing and applying bioinformatics pipelines tailored to specific needs. I have also gained considerable experience in Hi-C scaffolding and genome annotation. I view this continuous exchange of ideas as key to my growth as both a researcher and a contributing member of the broader scientific community.

DeepGOPlus is a bioinformatics tool designed for protein function prediction using deep learning. It combines convolutional neural networks with sequence-based features to predict Gene Ontology (GO) terms for proteins. This tool is particularly useful for large-scale genome annotation and proteome analysis, though a webserver is currently available for smaller requests. Frequently, genome annotation projects will produce tens of thousands of genes, and depending upon the model status of your chosen species, many of these genes may lack significant homology to anything available in current databases. This software is not a replacement for identifying ontology terms using interproscan, but an alternative for novel gene candidates.

Software used in this tutorial

  • Kulmanov, M., & Hoehndorf, R. (2020). DeepGOPlus: Improved protein function prediction from sequence. Bioinformatics, 36(2), 422–429. https://doi.org/10.1093/bioinformatics/btz595
  • Buchfink, B., Xie, C., & Huson, D. H. (2015). Fast and sensitive protein alignment using DIAMOND. Nature Methods, 12(1), 59–60. https://doi.org/10.1038/nmeth.3176
  • Github Repository

Installation of DeepGOPlus and prerequisite software

Create conda environment in python 3.7.9

  • create conda environment
  • activate conda environment
  • create python virtual environment
  • upgrade pip and setuptools
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

ml miniconda3
conda create -n deepgoplus_env python=3.7.9

#if you havent done this previously you'll need to run this to use conda
conda init
source ~/.bashrc

#activate the environment
conda activate deepgoplus_env

#create and actiavte a python virtual environment
python3 -m venv DeepGoPlus_pyenv
source DeepGoPlus_pyenv/bin/activate

#upgrade pip and setuptools
pip install --upgrade pip setuptools

Download the repository

  • Download the repository
  • install the requirements
1
2
3
4

git clone https://github.com/bio-ontology-research-group/deepgoplus.git
cd deepgoplus/
pip install -r requirements.txt
pip install deepgoplus

Download the metadata and training datasets

  • download datasets
  • copy metadata to data folder
1
2
3
4
5
6

wget http://deepgoplus.bio2vec.net/data/data.tar.gz

#automatically unpacks into data folder if within deepgoplus
tar -zxvf data.tar.gz

cp -rf metadata/ data/.

Install Diamond BLAST

  • download tar ball
  • add to PATH variable
1
2
3
4
5

wget http://github.com/bbuchfink/diamond/releases/download/v2.1.9/diamond-linux64.tar.gz
tar xzf diamond-linux64.tar.gz

#add to .bashrc and source or run in terminal for temporary access
export PATH="/work/gif3/masonbrink/Software/:$PATH"

Run DeepGoPlus analysis

  • Download some fun data from a plant parasitic nematode (Ditylenchus dipsaci)
  • Format the data
  • Run DeepGoPlus
  • Format the output
1
2
3
4
5
6
7
8
9
10
11
12
13

wget https://ftp.ebi.ac.uk/pub/databases/wormbase/parasite/releases/WBPS19/species/ditylenchus_dipsaci/PRJNA498219/ditylenchus_dipsaci.PRJNA498219.WBPS19.protein.fa.gz

gunzip ditylenchus_dipsaci.PRJNA498219.WBPS19.protein.fa.gz

# this will save formatting steps later, reducing the fasta header to just the basic name (first column).
less ditylenchus_dipsaci.PRJNA498219.WBPS19.protein.fa |awk '{print $1}' >Cleanditylenchus_dipsaci.PRJNA498219.WBPS19.protein.fa

# use the same threshold as the online database.  This allows all GO terms with a threshold of 0.3 to appear in the output. 
ml miniconda3; conda activate deepgoplus_env; source bin/activate ; deepgoplus -if Cleanditylenchus_dipsaci.PRJNA498219.WBPS19.protein.fa -dr data/ -of results_DDipsaci -t 0.3

# format the output into tabular form
awk 'NF>3' results_DDipsaci |sed 's/\t/#/1' |sed 's/\t/,/g' |sed 's/#/\t/g' >DeepGoPlusResults.tab
Gene ID GO Terms
jg20583 GO:0110165-0.355, GO:0005575-0.361
jg20588 GO:0110165-0.349, GO:0005575-0.357
jg16733 GO:0003674-0.567, GO:0005488-0.500, GO:0005515-0.377, GO:0005575-0.725, GO:0005622-0.613, GO:0005737-0.575, GO:0005886-0.331, GO:0007154-0.330, GO:0008150-0.727, GO:0009987-0.583, GO:0016020-0.499, GO:0023052-0.332, GO:0043226-0.400, GO:0043227-0.329, GO:0043229-0.359, GO:0043231-0.311, GO:0050789-0.481, GO:0050794-0.450, GO:0050896-0.393, GO:0051716-0.344, GO:0065007-0.494, GO:0071944-0.352, GO:0110165-0.720
jg20471 GO:0110165-0.387, GO:0005575-0.396
jg2136 GO:0003674-0.503, GO:0005488-0.369, GO:0005515-0.349, GO:0005575-0.788, GO:0005622-0.688, GO:0005737-0.635, GO:0005886-0.434, GO:0008104-0.443, GO:0008150-0.742, GO:0009987-0.687, GO:0016020-0.654, GO:0016043-0.382, GO:0032991-0.303, GO:0033036-0.444, GO:0033365-0.308, GO:0043226-0.566, GO:0043227-0.528, GO:0043229-0.555, GO:0043231-0.503, GO:0045184-0.435, GO:0051179-0.510, GO:0051234-0.500, GO:0051641-0.465, GO:0051668-0.362, GO:0070727-0.443, GO:0071840-0.387, GO:0071944-0.435, GO:0072657-0.359, GO:0090150-0.353, GO:0110165-0.783
jg12535 GO:0110165-0.345, GO:0005575-0.345
jg2135 GO:0110165-0.395, GO:0005575-0.397, GO:0005622-0.365, GO:0016020-0.309, GO:0043229-0.329, GO:0043226-0.337
jg2134 GO:0110165-0.369, GO:0005575-0.370, GO:0005622-0.343
jg12533 GO:0110165-0.360, GO:0005575-0.364, GO:0016020-0.329
jg14715 GO:0110165-0.370, GO:0005575-0.383, GO:0005622-0.305
jg21581 GO:0110165-0.380, GO:0005575-0.380
jg2133 GO:0003674-0.732, GO:0005215-0.637, GO:0005575-0.637, GO:0005622-0.327, GO:0005737-0.327, GO:0005886-0.602, GO:0006810-0.737, GO:0006811-0.616, GO:0006812-0.598, GO:0008150-0.948, GO:0008324-0.354, GO:0009987-0.771, GO:0015075-0.362, GO:0015095-0.319, GO:0015318-0.575, GO:0015693-0.555, GO:0016020-0.603, GO:0022857-0.637, GO:0022890-0.566, GO:0030001-0.589, GO:0034220-0.616, GO:0046873-0.340, GO:0050896-0.334, GO:0051179-0.737, GO:0051234-0.737, GO:0051716-0.307, GO:0055085-0.679, GO:0071944-0.602, GO:0098655-0.596, GO:0098660-0.607, GO:0098662-0.595, GO:0110165-0.637, GO:1903830-0.445
jg12534 GO:0110165-0.394, GO:0005575-0.396, GO:0005622-0.314

DeepGoPlus filtration

The raw output of DeepGoPlus gives a lot of broad terms to filter through if you are looking for specific details about one or a set of proteins. Here is an awk script that you can modify to your liking that will remove very broad GO terms.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

awk -F'\t' '  # Use tab as the field separator
{
    filtered_terms = "";  # Initialize an empty string to store filtered GO terms

    # Loop through columns 2 to NF (the last column)
    for (i = 2; i <= NF; i++) {
        split($i, parts, ":");  # Split each column into parts by the colon (":") separator
        go_id = parts[2];  # The second part (after the colon) is assigned to go_id, which represents the GO term ID

        # Exclude broad and less specific GO terms by checking if go_id matches any excluded terms
        if (go_id !~ /^GO:0008150$/ && go_id !~ /^GO:0003674$/ && go_id !~ /^GO:0005575$/ && \
            go_id !~ /^GO:0009987$/ && go_id !~ /^GO:0071840$/ && go_id !~ /^GO:0050896$/ && \
            go_id !~ /^GO:0005488$/ && go_id !~ /^GO:0003824$/ && go_id !~ /^GO:0060089$/ && \
            go_id !~ /^GO:0038023$/ && go_id !~ /^GO:0098772$/ && go_id !~ /^GO:0030246$/ && \
            go_id !~ /^GO:0032991$/ && go_id !~ /^GO:0044464$/ && go_id !~ /^GO:0044424$/ && \
            go_id !~ /^GO:0044422$/ && go_id !~ /^GO:0043231$/ && go_id !~ /^GO:0043229$/) {

            # If filtered_terms is not empty, add a tab to separate terms
            if (filtered_terms != "") {
                filtered_terms = filtered_terms "\t";
            }

            # Append the current GO term (column $i) to the filtered_terms string
            filtered_terms = filtered_terms $i;
        }
    }

    # If any filtered terms were found (filtered_terms is not empty), print the first column and the filtered GO terms
    if (filtered_terms != "") {
        print $1 "\t" filtered_terms;
    }
}' results_DDipsaci > filtered_output.txt

Here is a list of the current terms that are being eliminated from the results

GO Term Definition
GO:0008150 Biological process
GO:0003674 Molecular function
GO:0005575 Cellular component
GO:0009987 Cellular process
GO:0071840 Cellular component organization or biogenesis
GO:0050896 Response to stimulus
GO:0005488 Binding
GO:0003824 Catalytic activity
GO:0060089 Molecular transducer activity
GO:0038023 Signaling receptor activity
GO:0098772 Molecular function regulator
GO:0030246 Carbohydrate binding
GO:0032991 Macromolecular complex
GO:0044464 Cell part
GO:0044424 Intracellular part
GO:0044422 Organelle part
GO:0043231 Intracellular membrane-bounded organelle
GO:0043229 Intracellular organelle

Back to the Assembly and Annotation Index page