symcla: symbiont classifier

💾 Installation

Clone the symcla repository:

git clone https://github.com/NeLLi-team/symcla.git

cd symcla/
chmod u+x symcla

Create conda environment and install requirements:

conda create -c conda-forge -c bioconda --name symcla --file requirements.txt

💽 Setup data (run only once)

Run inside the symcla/ folder:

conda activate symcla

./symcla setup

---

🚀 Example run

conda activate symcla

cd symcla/

👷🏻‍♀️ Run the classifier

Run inside the symcla/ folder:

./symcla classify --genomedir data/test_genomes --savedir test_output --ncpus 32

To get help

./symcla classify --help

# Usage: symcla classify [OPTIONS]
#
# ╭─ Options ──────────────────────────────────────────────────────────────────────╮
# │ --genomedir                      TEXT     [default: input_genomes]             │
# │ --savedir                        TEXT     [default: output_symcla]             │
# │ --ncpus                          INTEGER  [default: 16]                        │
# │ --deltmp          --no-deltmp             [default: deltmp]                    │
# │ --help                                    Show this message and exit.          │
# ╰────────────────────────────────────────────────────────────────────────────────╯

🕺🏻 Results

Expected results from the test data:

taxon_oid	completeness_UNI56	features_gt0	features_ge20	features_ge100	symcla_score
IMGI2140918011	98.214	396	291	128	0.000
IMGI2645727657	100.000	287	197	70	-0.003
IMGI651324087	100.000	368	252	106	-0.009
IMGM3300027739_BIN74	64.286	310	234	95	0.001
SCISO2808607008	98.214	406	276	124	2.000
SDISOGCA_003484685.1	83.929	193	126	43	2.000
SHISO2654587767	98.214	423	309	134	2.000
SLISOGCF_900639865.1	100.000	569	429	234	0.999
SRISO640427127	92.857	296	197	103	2.000
SXGCA_000019745.1	98.214	353	259	106	0.126
SXGCA_902860225.1_Azoamicus_ciliaticola	91.071	117	83	36	1.055
SXISO642555114	96.429	333	243	108	1.995

🧐 Interpretation of results:

• completeness_UNI56: The percentage of 56 universal bacterial and archaeal marker genes found in the genome. We do not advise to trust any results <50%. Confidence in symbiont prediction increases with UNI56 completeness.
• features_gt01: Number of features found with a bitscore greater than 0. Confidence in symbiont prediction increases with more features found.
• features_ge20: Number of features found with a bitscore greater or equal than 20. Confidence in symbiont prediction increases with more features found.
• features_ge100: Number of features found with a bitscore greater or equal than 100. Confidence in symbiont prediction increases with more features found.
• symcla_score: after adjusting the classification thresholds based on thousands of experiments, we recommend the following values:
  • symcla_score <= 0.42: Free-living
  • 0.42 < symcla_score < 1.21: Symbiont;Host-associated
  • symcla_score >= 1.21: Symbiont;Intracellular

🤖 Note: by design symcla minimizes the rate of false positives for symbionts, at the expense of increased false negatives (i.e. some Symbiont;Host-associated might still get a symcla score lower than 0.42, and some Symbiont;Intracellular might still get a symcla score lower than 1.21).

🐳 symcla container

Apptainer

⚠️⚠️⚠️ Note: The paths to the input genomes and output directory must be absolute. ⚠️⚠️⚠️

apptainer pull \
        docker://docker.io/jvillada/symcla:latest


apptainer run \
        --pwd /usr/src/symcla \
        docker://docker.io/jvillada/symcla:latest \
        symcla \
        classify \
        --genomedir absolute_path_to_dir_with_faa_files \
        --savedir absolute_path_to_output_dir \
        --ncpus 16