tyPPing and other methods for detecting phage-plasmids

For more information, see: "Efficient detection and typing of phage-plasmids"

Here you can find the documentation, scripts, and data necessary for the identification of phage-plasmids (P-Ps):

Data and scripts required to reproduce analysis and figures of our study are in 'Publication_related_data'.

tyPPing

tyPPing is a user-friendly, fast and accurate method to detect P-Ps. Currently it finds P-Ps of the type AB_1, P1_1, P1_2, N15, SSU5, pMT1, pCAV, pSLy3, pKpn, and cp32. It uses protein profiles to search sequences for patterns (frequency and compositional sets) of conserved P-P proteins. If a match also fits the typical size range, it is predicted as a P-P with a distinct confidence.

MM-GRC

MM-GRC (multi-model gene repertoire clustering) is our first method to classify P-Ps (see PMID: 33590101). It is an integrated approach combining functional annotation (with phage- and plasmid-specific HMM profiles), machine learning (random forest) models, and which was complemented with an exhaustive literature review. It relies on the gene repertoire relatedness to type P-Ps, and detects various types including diverse communities and unrelated putative P-Ps (singletons).

geNomad_vConTACT2

Here we describe how to use geNomad and vConTACT v2 for detecting and typing P-Ps.

geNomad classifies nucleotide sequences as phages, integrated prophages, or plasmids. In our study, we used it to analyze sequences of a plasmid database, cases classed as phages were considered as potential P-Ps.

vConTACT v2 is clusters viral genomes with a reference dataset using their shared gene content. Here, we used it to group the putative P-Ps identified by geNomad with 1416 P-Ps that we typed in previous work.

Publication_related_data

Scripts, data, and supplementary materials to reproduce the figures and analyses presented in the publication.

Zenodo repository

Further files are available at the Zenodo repository.

• tyPPing_signature_profiles.hmm – 763 HMM profiles (concatenated) specific to the 10 P-P types. Used for protein-to-profile comparison (needed for tyPPing).

• phage.hmm – phage-specific HMM profiles (required for MM-GRC).

• models/ – random forest models trained to detect P-Ps in plasmid datasets (used by MM-GRC).

• g2g_plot_tables/ and tyPPing_criteria_tables/ contain the tables required to produce figures of our study using the scripts all_g_to_g_plots_filtered.R and figures_methods.R (in Publication_related_data/).

• draft_genomes_analysis/ – 12 complete P-P genomes that we detected in 9 draft genomes of carbapenem-resistant Enterobacteriales species. We used a modified version of tyPPing, tyPPing_for_draft_genomes.R, on drafts assembled by short and long reads, and on hybrid assemblies.

References

• MM-GRC
• HMMER
• MMseqs2
• geNomad
• vConTACTv2
• PHROGs
• Migale

Citing tyPPing

If you like tyPPing and you use it for your work, please cite: "Efficient detection and typing of phage-plasmids"