WO2025054593A2 - Désextinction moléculaire d'antibiotiques activés par apprentissage profond - Google Patents

Désextinction moléculaire d'antibiotiques activés par apprentissage profond Download PDF

Info

Publication number
WO2025054593A2
WO2025054593A2 PCT/US2024/045821 US2024045821W WO2025054593A2 WO 2025054593 A2 WO2025054593 A2 WO 2025054593A2 US 2024045821 W US2024045821 W US 2024045821W WO 2025054593 A2 WO2025054593 A2 WO 2025054593A2
Authority
WO
WIPO (PCT)
Prior art keywords
peptides
apex
peptide
antimicrobial
deep learning
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
PCT/US2024/045821
Other languages
English (en)
Other versions
WO2025054593A3 (fr
Inventor
César de la Fuente-Nunez
Marcelo DER TOROSSIAN TORRES
Fangping WAN
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Pennsylvania Penn
Original Assignee
University of Pennsylvania Penn
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by University of Pennsylvania Penn filed Critical University of Pennsylvania Penn
Publication of WO2025054593A2 publication Critical patent/WO2025054593A2/fr
Publication of WO2025054593A3 publication Critical patent/WO2025054593A3/fr
Anticipated expiration legal-status Critical
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
    • G16B40/20—Supervised data analysis
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61P—SPECIFIC THERAPEUTIC ACTIVITY OF CHEMICAL COMPOUNDS OR MEDICINAL PREPARATIONS
    • A61P31/00—Antiinfectives, i.e. antibiotics, antiseptics, chemotherapeutics
    • A61P31/04—Antibacterial agents
    • C—CHEMISTRY; METALLURGY
    • C07—ORGANIC CHEMISTRY
    • C07K—PEPTIDES
    • C07K14/00—Peptides having more than 20 amino acids; Gastrins; Somatostatins; Melanotropins; Derivatives thereof
    • C—CHEMISTRY; METALLURGY
    • C07—ORGANIC CHEMISTRY
    • C07K—PEPTIDES
    • C07K7/00—Peptides having 5 to 20 amino acids in a fully defined sequence; Derivatives thereof
    • C07K7/04—Linear peptides containing only normal peptide links
    • C07K7/08—Linear peptides containing only normal peptide links having 12 to 20 amino acids
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61K—PREPARATIONS FOR MEDICAL, DENTAL OR TOILETRY PURPOSES
    • A61K38/00—Medicinal preparations containing peptides
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B35/00—ICT specially adapted for in silico combinatorial libraries of nucleic acids, proteins or peptides
    • G16B35/20—Screening of libraries

Definitions

  • TECHNICAL FIELD [0003] The present disclosure pertains to identification, synthesis, and use antibiotic peptides.
  • BACKGROUND [0004] With antimicrobial resistant (AMR) infections causing approximately 1.27 million deaths annually worldwide, and projections indicating a potential 10 million annual fatalities by 2050 3 in the absence of effective new drugs, urgent measures are required to combat antibiotic resistance. Furthermore, according to the World Health Organization, by 2030, around 24 million individuals could face extreme poverty due to the high cost of treating these infections 3 . [0005] Computational approaches have been developed for the design and discovery of peptide antibiotics 4 .
  • ML machine learning
  • antimicrobial peptides deriving from extinct proteomes using a multitask deep learning approach.
  • antimicrobial peptides and methods of treating an antimicrobial infection comprising contacting the infection with the present antimicrobial peptides.
  • FIG.1 depicts how molecular de-extinction of antibiotics from proteomes was accomplished using deep learning. All available proteomes of extinct organisms were mined by APEX, our deep learning algorithm. Amino acid sequences ranging from 8 to 50 amino acid residues within proteins from extinct organisms were inputted into a multitask deep learning model that trained on both public and in-house peptide data to evaluate the potential antimicrobial activity.
  • FIGS.2A and 2B illustrate how APEX accurately identified antimicrobials in extinct organisms.
  • FIG.2A provides a radar chart showing R-squared correlation in terms of species-specific antimicrobial activity prediction on an independent dataset (a held-out subset from our in-house peptide dataset) for various machine learning (ML) models. The radius reflects the R-squared value for each of the models.
  • FIG.2B depicts the mean of species-wise Pearson correlation of log 2 -transformed MICs between values obtained experimentally and predicted by various ML models. Evaluated dataset: 69 peptides were synthesized and tested. 24-10510 (103241.007081) [0010]
  • FIGS.3A-3B illustrate antimicrobials identified by APEX in extinct organisms and their composition and physicochemical properties.
  • FIG.3A provides a phylogenetic tree showing the extinct organisms scanned by APEX. Circular bars denote the log 10 -transformed average active (red) and inactive (blue) encrypted peptides discovered by APEX. A peptide was considered active when its predicted median MIC against the bacterial strains tested was ⁇ 80 ⁇ mol L -1 . The values were normalized by the number of proteins per organism scanned. The organisms whose encrypted peptides were selected for validation are highlighted in bold type. Extinct organisms that presented active encrypted peptides (EPs) validated experimentally are indicated by a light red square and, within that group, those organisms encoding extinct sequences absent in extant organisms are highlighted with a dark red square.
  • EPs active encrypted peptides
  • FIG.3B shows the amino acid frequency in AEPs and MEPs compared with known AMPs from the DBAASP database.
  • AEPs present a higher frequency of the basic residue K, the aliphatic residue V, and uncharged polar residues (M, Q, and T) than MEPs.
  • FIGS.3C and 3D show the distribution of two physicochemical properties for peptides with predicted antimicrobial activity (AEPs and MEPs) and AMPs from DBAASP: net charge in FIG.3C; and normalized hydrophobicity in FIG.3D. Net charge directly influences the initial electrostatic interactions between the peptide and negatively charged bacterial membranes, and hydrophobicity directly influences the interactions of the peptide with lipids in the membrane bilayers.
  • Encrypted peptides from extinct organisms are slightly less hydrophobic, and similarly have a net positive charge, when compared with encrypted peptides from the modern human proteome 15 or peptides from DBAASP.
  • Statistical significance in c and d was determined using two-tailed t-tests followed by Mann-Whitney test; p values are shown in the graph. The solid line inside each box represents the mean value obtained for each group.
  • FIGS.4A and 4B pertain to mntimicrobial activity profiles of sequences from the proteomes of extinct organisms.
  • FIG.4A provides a heat map of the antimicrobial activities ( ⁇ mol L ⁇ 1 ) of the active encrypted peptides from extinct organisms against 11 clinically relevant pathogens, including strains resistant to conventional antibiotics. Briefly, 10 6 bacterial cells and serially diluted encrypted peptides (0-128 ⁇ mol L ⁇ 1 ) were incubated at 37 °C. One day post-treatment, the optical density at 600 nm was measured in a microplate reader to evaluate bacterial growth in the presence of the encrypted peptides from extinct organisms. MIC values in the heat map are the arithmetic 24-10510 (103241.007081) mean of the replicates in each condition.
  • FIG.4B provides examples of active archaic encrypted peptides (AEPs) and modern encrypted peptides (MEPs) from various extinct organisms, their parent protein, and their activity profile against ESKAPE pathogens (Enterococcus spp., S. aureus, K. pneumoniae, A. baumannii, P. aeruginosa, E. coli).
  • Antimicrobial activity is expressed as the MIC ( ⁇ mol L ⁇ 1 ) and activity bars are presented as -log 2 MIC 13 .
  • the data for the assays in a are the mean and the experiments were performed in three independent replicates.
  • AEPs in a are indicated by an asterisk (*).
  • FIGS.5A-5E illustrate the antimicrobial activity, mechanism of action, and synergy of antimicrobials from the proteomes of extinct organisms.
  • FIG.5A provides pan-bacterial Pearson and Spearman correlations of log 2 -transformed MICs between experimentally validated values and values predicted by APEX.
  • FIG.5B provides a comparison between the hit rates of APEX and the scoring function previously described by Torres et al. 15 to detect encrypted peptides in the modern human proteome.
  • FIG.5C shows cytoplasmic membrane depolarization by five encrypted peptides from extinct organisms. The A. baumannii membrane was more strongly depolarized by the encrypted peptides than by the antibiotic polymyxin B.
  • FIG.5D provides NPN permeabilization assays showing the effect of two encrypted peptides from extinct organisms on the outer membrane of A. baumannii. Higher permeability was observed with the encrypted peptides than with the antibiotic polymyxin B.
  • FIG.5E is a heat map showing interactions between encrypted peptides identified by APEX, expressed as the fractional inhibitory concentration index (FICI). Most of the tested encrypted peptide pairs from extinct organisms either synergized or had an additive effect against A. baumannii and P. aeruginosa; the latter was only tested against the peptide pair composed of equusin-1 and equusin-2 shown in the last row of the heatmap.
  • FICI fractional inhibitory concentration index
  • FIGS.6A-6D pertain to studies of the anti-infective activity of encrypted peptides in animal models.
  • FIG.6B shows how the encrypted peptides mammuthusin-2 (Mammuthus primigenius), hydrodamin-1 (Hydrodamalis gigas), megalocerin-1 (Megalocerus sp.), elephasin-2 (Elephas antiquus), and mylodonin- 24-10510 (103241.007081) 2 (Mylodon darwinii), administered at their MIC in a single dose inhibited the proliferation of the infection for up to four days after treatment compared to the untreated control group. Elephasin-2 and mylodonin-2 cleared the infection in some of the mice, with activity comparable to that of the antibiotic used as control, polymyxin B.
  • FIG.6D shows how, two days after intraperitoneal injection, mylodonin-2 at its MIC reduced A. baumannii ATCC19606 infection as much as polymyxin B, compared to the untreated control group. Four days post-treatment, mammuthusin-2 and elephasin-2 showed the same level of activity as polymyxin B.
  • FIG.7 provides a schematic illustration of APEX.
  • FIG.8. shows R-squared scores of various ML models on cross- validation (CV) set.
  • FIG.9 shows Pearson correlation scores of various ML models on cross- validation (CV) set.
  • FIG.10 shows Spearman correlation scores of various ML models on cross-validation (CV) set.
  • FIG.11 depicts R-squared scores of various ML models on cross- validation (CV) set.
  • FIG.12 shows Pearson correlation scores of various ML models on CV set.
  • FIG.13 provides Spearman correlation scores of various ML models on CV set.
  • FIG.14 shows the relationship between R-squared and the number of APEX models used in ensemble learning on the CV set. 24-10510 (103241.007081)
  • FIG.15 depicts the relationship between Pearson correlation and the number of APEX models used in ensemble learning on the CV set.
  • FIG.22 provides R-squared scores of various ML models on an independent set.
  • FIG.23 provides Pearson correlation scores of various ML models on an independent set.
  • FIG.24 shows Spearman correlation scores of various ML models on an independent set.
  • FIG.25 provides R-squared scores of ensemble APEX v2 and v1 on an independent set.
  • FIG.26 shows Pearson correlation scores of ensemble APEX v2 and v1 on an independent set.
  • FIG.27 provides Spearman correlation scores of ensemble APEX v2 and v1 on an independent set.
  • FIG.28 shows Pearson correlation scores of ensemble APEX v2 and the scoring function used to identify modern human encrypted peptides.
  • FIG.29 provides Spearman correlation scores of ensemble APEX v2 and the scoring function used to identify modern human encrypted peptides.
  • FIG.30 shows Pearson correlation scores of various ML models used to identify modern human encrypted peptides. 24-10510 (103241.007081)
  • FIG.31 provides Spearman correlation scores of various ML models used to identify modern human encrypted peptides.
  • FIG.32 shows sequence space exploration using a similarity matrix.
  • FIG.33 depicts the relative abundance of the amino acid content of encrypted peptides (EPs) from the modern human proteome identified by APEX (top) and the scoring function (bottom).
  • FIG.34 shows the relative abundance of the amino acid content of encrypted peptides (EPs) identified by APEX from the proteomes of extinct organisms (top) compared to known AMPs from DBAASP (bottom).
  • FIG.35 shows the relative abundance of the amino acid content of archaic encrypted peptides (AEPs) identified by APEX from the proteomes of extinct organisms (top) compared to known AMPs from DBAASP (bottom).
  • FIG.36 depicts the relative abundance of the amino acid content of MEPs identified by APEX from the proteomes of extinct organisms compared to known AMPs from DBAASP.
  • FIG.37 shows the relative abundance of the amino acid content of AEPs and MEPs identified by APEX from the proteomes of extinct organisms.
  • FIGS.38A-F depict physicochemical features of AEPs and MEPs identified by APEX in extinct organisms compared to AMPs from DBAASP.
  • FIGS.39A-D concern the secondary structure of active EPs predicted by the scoring function and APEX in helical inducer medium.
  • FIG.40 shows the antimicrobial activity of encrypted peptides from extinct organisms predicted by the scoring function.
  • FIG.41 illustrates sequence space exploration using a similarity matrix containing the 69 encrypted peptides discovered by APEX selected for further experimental validation compared to peptide sequences from DBAASP.
  • FIG.42 shows predicted vs. experimental MICs for A. baumannii ATCC 19606 of the encrypted peptides identified by APEX.
  • FIG.43 shows predicted vs. experimental MICs for E. coli AIC221 of the encrypted peptides identified by APEX.
  • FIG.44 shows predicted vs experimental MICs for E. coli AIC222 of the encrypted peptides identified by APEX. 24-10510 (103241.007081)
  • FIG.45 shows predicted vs.
  • FIG.46 shows predicted vs. experimental MICs for K. pneumoniae ATCC 13883 of the encrypted peptides identified by APEX.
  • FIG.47 depicts predicted vs. experimental MICs for P. aeruginosa PA14 of the encrypted peptides identified by APEX.
  • FIG.48 shows predicted vs. experimental MICs for P. aeruginosa PAO1 of the encrypted peptides identified by APEX.
  • FIG.49 provides predicted vs. experimental MICs for methicillin- resistant S.
  • FIG.50 shows predicted vs. experimental MICs for S. aureus ATCC 12600 of the encrypted peptides identified by APEX.
  • FIG.51 provides predicted vs. experimental MICs for vancomycin- resistant E. faecalis ATCC 700802 of the encrypted peptides identified by APEX.
  • FIG.52 provides predicted vs. experimental MICs for vancomycin- resistant E. faecium ATCC 700221 of the encrypted peptides identified by APEX.
  • FIGS.53A-D pertain to cytoplasmic membrane depolarization of A. baumannii and P.
  • FIGS.54A-D pertain to outer membrane permeabilization of A. baumannii and P. aeruginosa cell membranes caused by encrypted peptides from extinct organisms.
  • FIG.55 depicts the synergy between peptide molecules from extinct organisms.
  • FIG.56 shows the results of assays of the resistance to proteolytic degradation.
  • FIG.57 depicts the results of weight change monitoring in the skin abscess mouse model infected with A. baumannii.
  • FIG.58 depicts the results of weight change monitoring in the thigh mouse model infected with A. baumannii.
  • this disclosure proposes molecular de-extinction as a framework for drug discovery, aiming to address the urgent global health issue of antimicrobial-resistant (AMR) infections.
  • AMR antimicrobial-resistant
  • This disclosure introduces Antibiotic Peptide de-Extinction (APEX, Fig. 1), a new multitask deep learning (DL) approach.
  • APEX Antibiotic Peptide de-Extinction
  • DL deep learning
  • the inventors systematically mined all available proteomes of extinct organisms (the “extinctome”) to discover potential antimicrobial peptides. This effort led to the identification of 37,176 EPs with predicted antibiotic activity (Data S1).
  • AEPs archaic EPs
  • MEPs modern EPs
  • antimicrobial peptides having an amino acid sequence of any of SEQ ID NOs:1-41.
  • compositions comprising an antimicrobial peptide having an amino acid sequence of any of SEQ ID NOs:1-41, or an antimicrobial peptide that has been identified using a presently disclosed method, and a pharmaceutically acceptable carrier, diluent, or excipient.
  • the compositions may include two or more peptides of SEQ ID NOs:1-41 or two more antimicrobial peptides that have been identified using a presently disclosed method.
  • the present disclosure also provides methods treating an antimicrobial infection comprising contacting the infection with a therapeutically effective amount of an antimicrobial peptide that has been identified according to a disclosed method, such as an antimicrobial peptide of any one of SEQ ID NOs:1-41, or a composition comprising a therapeutically effective amount of an antimicrobial peptide that has been identified according to a disclosed method, such as an antimicrobial peptide of any one of SEQ ID NOs:1-41.
  • the phrase “therapeutically effective amount” refers to the amount of active agent (here, the antimicrobial peptide) that elicits the biological or medicinal response that is being sought in a tissue, system, animal, individual or human by a researcher, veterinarian, medical doctor or other clinician, which includes one or more of the following: 24-10510 (103241.007081) (1) at least partially preventing the disease or condition or a symptom thereof; for example, preventing a disease, condition or disorder in an individual who may be predisposed to the disease, condition or disorder but does not yet experience or display the pathology or symptomatology of the disease; (2) inhibiting the disease or condition; for example, inhibiting a disease, condition or disorder in an individual who is experiencing or displaying the pathology or symptomatology of the disease, condition or disorder (i.e., including arresting further development of the pathology and/or symptomatology); and (3) at least partially ameliorating the disease or condition; for example, ameliorating a disease, condition or disorder in an individual who is experiencing
  • the antimicrobial peptides that are administered, contacted with a biofilm, or included in a composition to the present disclosure may be provided in a composition that is formulated for any type of administration.
  • the compositions may be formulated for administration orally, topically, parenterally, enterally, or by inhalation (e.g., intranasally).
  • the active agent may be formulated for neat administration, or in combination with conventional pharmaceutical carriers, diluents, or excipients, which may be liquid or solid.
  • the applicable solid carrier, diluent, or excipient may function as, among other things, a binder, disintegrant, filler, lubricant, glidant, compression aid, processing aid, color, sweetener, preservative, suspensing/dispersing agent, tablet-disintegrating agent, encapsulating material, film former or coating, flavoring agent, or printing ink.
  • a binder disintegrant, filler, lubricant, glidant, compression aid, processing aid, color, sweetener, preservative, suspensing/dispersing agent, tablet-disintegrating agent, encapsulating material, film former or coating, flavoring agent, or printing ink.
  • Any material used in preparing any dosage unit form is preferably pharmaceutically pure and substantially non-toxic in the amounts employed.
  • the active agent may be incorporated into sustained-release preparations and formulations.
  • Administration in this respect includes administration by, inter alia, the following routes: intravenous, intramuscular, subcutaneous, intraocular, intrasynovial, transepithelial including transdermal, ophthalmic, sublingual and buccal; topically including ophthalmic, dermal, ocular, rectal and nasal inhalation via insufflation, aerosol, and rectal systemic.
  • the carrier, diluent, or excipient may be a finely divided solid that is in admixture with the finely divided active ingredient.
  • the active 24-10510 (103241.007081) ingredient is mixed with a carrier, diluent or excipient having the necessary compression properties in suitable proportions and compacted in the shape and size desired.
  • the active compound may be incorporated with the carrier, diluent, or excipient and used in the form of ingestible tablets, buccal tablets, troches, capsules, elixirs, suspensions, syrups, wafers, and the like.
  • the amount of active agent(s) in such therapeutically useful compositions is preferably such that a suitable dosage will be obtained.
  • Liquid carriers, diluents, or excipients may be used in preparing solutions, suspensions, emulsions, syrups, elixirs, and the like.
  • the active ingredient of this invention can be dissolved or suspended in a pharmaceutically acceptable liquid such as water, an organic solvent, a mixture of both, or pharmaceutically acceptable oils or fat.
  • the liquid carrier, excipient, or diluent can contain other suitable pharmaceutical additives such as solubilizers, emulsifiers, buffers, preservatives, sweeteners, flavoring agents, suspending agents, thickening agents, colors, viscosity regulators, stabilizers, or osmo-regulators.
  • suitable pharmaceutical additives such as solubilizers, emulsifiers, buffers, preservatives, sweeteners, flavoring agents, suspending agents, thickening agents, colors, viscosity regulators, stabilizers, or osmo-regulators.
  • Suitable solid carriers, diluents, and excipients may include, for example, calcium phosphate, silicon dioxide, magnesium stearate, talc, sugars, lactose, dextrin, starch, gelatin, cellulose, methyl cellulose, ethylcellulose, sodium carboxymethyl cellulose, microcrystalline cellulose, polyvinylpyrrolidine, low melting waxes, ion exchange resins, croscarmellose carbon, acacia, pregelatinized starch, crospovidone, HPMC, povidone, titanium dioxide, polycrystalline cellulose, aluminum methahydroxide, agar-agar, tragacanth, or mixtures thereof.
  • liquid carriers, diluents and excipients for example, for oral, topical, or parenteral administration, include water (particularly containing additives as above, e.g. cellulose derivatives, preferably sodium carboxymethyl cellulose solution), alcohols (including monohydric alcohols and polyhydric alcohols, e.g. glycols) and their derivatives, and oils (e.g. fractionated coconut oil and arachis oil), or mixtures thereof.
  • the carrier, diluent, or excipient can also be an oily ester such as ethyl oleate and isopropyl myristate.
  • sterile liquid carriers diluents, or excipients, which are used in sterile liquid form compositions for parenteral administration.
  • Solutions of the active agents can be prepared in water 24-10510 (103241.007081) suitably mixed with a surfactant, such as hydroxypropylcellulose.
  • a dispersion can also be prepared in glycerol, liquid polyethylene glycols, and mixtures thereof and in oils. Under ordinary conditions of storage and use, these preparations may contain a preservative to prevent the growth of microorganisms.
  • the proper fluidity can be maintained, for example, by the use of a coating, such as lecithin, by the maintenance of the required particle size in the case of a dispersion, and by the use of surfactants.
  • a coating such as lecithin
  • surfactants for example, sodium bicarbonate, sodium bicarbonate, sodium bicarbonate, sodium bicarbonate, sodium bicarbonate, sodium bicarbonate, sodium bicarbonate, sodium bicarbonate, sodium sulfate, sodium sulfate, sodium sulfate, sodium sulfate, sodium sorbic acid, thimerosal and the like.
  • the antimicrobial peptides themselves may be sufficient to prevent contamination by microorganisms.
  • administration may be in the amount of about 0.1 mg/day, about 0.5 mg/day, about 1.0 mg/day, about 5 mg/day, about 10 mg/day, about 20 mg/day, about 50 mg/day, about 100 mg/day, about 200 mg/day, about 250 mg/day, about 300 mg/day, or about 500 mg/day.
  • methods comprising contacting a biofilm with an effective amount of an antimicrobial peptide that has been identified according to a presently disclosed method.
  • the antimicrobial peptide comprises one or more of SEQ ID NOs: 1-41.
  • microbes against which the present antimicrobial peptides are effective may be, for example, any unicellular organism, such as gram-negative bacteria, gram-positive bacteria, protozoa, viruses, bacteriophages, and archaea.
  • the present peptides can have an antimicrobial effect with respect to any such microbe.
  • bacteria against which the present compounds are effective to cause reduction in numbers include gram positive bacteria and gram negative bacteria, for example, Salmonella enterica, Listeria monocytogenes, Escherichia coli, Clostridium botulinum, Clostridium difficile, Campylobacter, Bacillus cereus, Vibrio parahaemolyticus, Vibrio cholerae, Vibrio vulnificus, Staphylococcus aureus, Yersinia enterocolitica, Shigella, Moraxella spp., Helicobacter, Stenotrophomonas, Bdellovibrio, Legionella spp.
  • Salmonella enterica Listeria monocytogenes
  • Escherichia coli Clostridium botulinum
  • Clostridium difficile Clostridium difficile
  • Campylobacter Bacillus cereus
  • Vibrio parahaemolyticus Vibrio cholerae
  • Vibrio vulnificus Vibrio vulnificus
  • Neisseria gonorrhoeae Neisseria meningitidis
  • Haemophilus influenzae Acinetobacter baumannii
  • Klebsiella pneumoniae Pseudomonas aeruginosa
  • Proteus mirabilis Enterobacter cloacae
  • Enterococcus faecium Serratia marcescens
  • Helicobacter pylori Salmonella enteritidis
  • Salmonella typhi and combinations thereof.
  • Salmonella enterica serovars that can be reduced using the compounds of the disclosure include, for example, Salmonella enteriditis, Salmonella typhimurium, Salmonella poona, Salmonella heidelberg, and Salmonella anatum.
  • Exemplary viruses against which the present peptides are effective to cause reduction in numbers include coronaviruses, rhinoviruses, and influenza viruses.
  • the mining includes setting a selectivity scoring method for selecting and filtering candidate antimicrobial peptides.
  • the present disclosure also pertains to methods of identifying an antimicrobial peptide comprising training and using a deep learning model according to steps disclosed in the present specification and figures. [0095] Also provided are methods of identifying an antimicrobial peptide comprising training and using a deep learning model according to steps depicted in FIG.1. Examples [0096] The present invention is further defined in the following Examples. It should be understood that the examples, while indicating preferred embodiments of the invention, are given by way of illustration only, and should not be construed as limiting the appended claims.
  • APEX utilizes an encoder neural network, combining recurrent and attention neural networks (FIG.7), to extract hidden features from peptide sequences.
  • the encoder neural network was then coupled with multiple downstream neural networks to predict antimicrobial activity according to the peptide source (i.e., from in-house or public datasets).
  • the present inventors randomly split the inventors’ in-house dataset into a cross validation (CV) set and an independent set, consisting of 790 and 198 peptides, respectively. Five-fold CV was first used to tune the hyperparameters on 24-10510 (103241.007081) the CV set, while the independent set was used to evaluate the final prediction performance of ML models trained on the CV set with determined hyperparameters. [0099] To compare the performance of the inventors’ deep learning approach with simple ML predictors, the present inventors implemented several baseline ML models, including elastic net, linear support vector regression, extra-trees regressor, random forest, and gradient boosting decision tree, and trained and evaluated them on the same datasets.
  • APEX involved the following multitask training steps: (1) using a single FCNN to simultaneously predict peptide antimicrobial activity for the 34 strains tested; (2) 24-10510 (103241.007081) augmenting the training data by incorporating another FCNN to predict whether peptides from public databases (either AMPs or non-AMPs) were antimicrobial; and (3) the present inventors imposed a multitask training constraint on the learnable weights of the last layer in the species-specific antimicrobial prediction FCNN. Briefly, this last constraint encouraged the model to give similar prediction results for similar bacteria (defined by having shorter phylogenetic distance from each other).
  • ensemble APEX v1 achieved an R-squared value of 0.520, a Pearson correlation of 0.706, and a Spearman correlation of 0.582 on average (FIG.2a, FIGS.16-20 and Supplementary Tables 12-14), outperforming all baseline ML methods.
  • a single ML model may be trained with different random seeds. The present inventors averaged the prediction results from all model copies to counter the potential stochastic behavior caused by the choice of random seeds.
  • the present inventors trained five copies with different random seeds and created a second ensemble learning version (ensemble APEX v2) with 40 APEX 24-10510 (103241.007081) models (i.e., eight APEX models ⁇ five copies).
  • This ensemble learning approach increased the prediction performance to 0.546, 0.728 and 0.607 in terms of R-squared, Pearson correlation, and Spearman correlation, respectively (FIG.2a, FIGS.20-27 and Supplementary Tables 12-14).
  • the present inventors then tested APEX’s predictive power compared to that of a scoring function used previously to discover encrypted peptide antibiotics in the modern human proteome 15 .
  • the scoring function 18 uses hydrophobicity and net charge to compute a predictive score of antimicrobial potential. Since the 56 human peptide antibiotics validated experimentally for antimicrobial activity in the inventors’ previous work 15 are part of the inventors’ in-house dataset, the present inventors used them here as the test set and used the rest of the inventors’ in-house dataset (932 peptides) for model training and selection. For the dataset consisting of the 56 validated human EPs, the ensemble APEX v2 achieved highest values for Pearson and Spearman correlations in most cases (FIGS.28-31 and Supplementary Tables 15-16).
  • APEX outperformed all baseline ML methods.
  • APEX achieved the highest Pearson correlation for MIC prediction (FIG.2b, more details on In vitro antimicrobial activity of antibiotic molecules from extinct organisms).
  • the inventors used its sequence similarities to all peptides (i.e., DBAASP peptides and 37,176 EPs) as its feature representation.
  • the inventors used the uniform manifold approximation and projection (UMAP) 20 technique to reduce the dimension of the feature representations to a bidimensional (2D) space (FIG.32). While DBAASP peptides mostly fell within the central area of the UMAP-derived 2D space, molecules identified by APEX had a much wider spread (FIG.32), forming multiple distinct clusters that were not covered by DBAASP.
  • APEX archaic encrypted peptides
  • molecules identified by APEX presented lower cysteine, aspartic acid, and glycine content compared to AMPs from DBAASP (FIG.34).
  • Peptides derived from proteins of extinct organisms also had a lower asparagine and higher methionine and glutamine content compared to AMPs from DBAASP (FIG.35).
  • the MEPs identified by APEX had a much lower alanine, proline, and tryptophan content but a much higher isoleucine, leucine, asparagine, and serine content than peptides within DBAASP (FIG.36).
  • the inventors obtained the predicted normalized hydrophobic moment (FIG.38e) and isoelectric point of AEPs (FIG.38f), which presented a low range of normalized hydrophobic moment (0- 0.6) and clustered within a short isoelectric point range (9.5 to 13). These values, found for sequences in extinct organisms (AEPs), overlapped with those obtained for sequences in extinct and extant organisms (MEPs) as well as in AMPs from DBAASP (FIG.38e-f). The values aligned with the lower abundance of acidic residues compared to basic ones, particularly lysine, within AEPs (FIG.3c-d).
  • AEPs identified by APEX represent a distinct family of peptides with a higher abundance of uncharged polar residues and increased aliphatic content (particularly isoleucine and leucine) with respect to other classes of peptide antibiotics, including other EPs 15,22,23 and AMPs 24 .
  • AEPs Like the AMPs in these families but unlike previously described EPs 15 (FIG.33) or conventional AMPs (FIG.34), AEPs have a high abundance of uncharged polar residues 21 . Leucine and isoleucine, in particular, are structurally important: the stiffness of these bulky branched residues limits the internal flexibility of the peptide, whereas other aliphatic residues favor specific foldamers during the folding process 25 . The difference between the amino acid composition of known AMPs and that of AEPs and MEPs results in significantly different physicochemical features (FIG.3c-d and FIG.38), reaffirming that the EPs identified by APEX are different from known antimicrobial peptides.
  • Example 2 In vitro antimicrobial activity of antibiotic molecules from extinct organisms [00112] To further validate APEX’s predictive power in identifying active encrypted peptide sequences from extinct organisms, the inventors synthesized and tested two non-overlapping sets of peptides: (i) 49 EPs predicted by a scoring function 15 (FIG. 40) and (ii) 69 EPs predicted by APEX (FIG.3a) and found in 98 extinct species. While the 49 EPs predicted by the scoring function were found in both extinct and extant 24-10510 (103241.007081) organisms (i.e., all were classified as MEPs), the APEX-predicted sequences included many that were unique to extinct organisms (20 AEPs and 49 MEPs).
  • APEX was built to predict species-specific antimicrobial activities and 69 EPs were selected based on multiple selection criteria. Specifically, the inventors ranked the 10,311,899 sequences derived from the extinct proteomes by median antimicrobial activities (i.e., broad- spectrum activity), or selectivity against Gram-positive or Gram-negative pathogens. For each ranked list, the inventors used the following criteria to filter out compounds: (i) length not ranging from 8 to 30 amino acid residues, (ii) sequences that are present in the inventors’ in-house dataset, (iii) with high sequence similarity to known AMPs from public databases, and (iv) EPs that are present in the modern human proteome.
  • the inventors synthesized the 21 AEPs and 48 MEPs identified by APEX from extinct organisms and experimentally determined their MICs for 11 clinically relevant bacterial pathogens (seven Gram-negatives and four Gram-positives), ten of which are on the ESKAPEE pathogen list 17 (FIG.4a-b).
  • the name of each source organism was used as the basis for the inventors’ molecular de-extinction nomenclature.
  • All experimentally determined MICs (log 2 transformed) were compared to predictions generated by APEX, yielding Pearson and Spearman correlation values of 0.448 and 0.404, respectively (FIG.5a), underscoring APEX’s substantial predictive power.
  • APEX showed a high predicted versus experimentally validated activity correlation (Pearson correlation >0.3) for A. baumannii ATCC 19606, Escherichia coli strains AIC221, AIC222 (a colistin-resistant strain), and ATCC 11775, P. aeruginosa strains PAO1 and PA14, and E. faecium ATCC 700221 (a vancomycin-resistant strain). All correlation results obtained for the 11 experimentally validated strains are shown in FIGS.42-52.
  • anomalopterin-1 a peptide originating from the extinct moa species Anomalopteryx didiformis. This peptide is a fragment of the dynein axonemal heavy chain 3, which forms part of the microtubule- associated motor protein complex.
  • Mylodonin-2 derived from the extinct South American giant sloth Mylodon darwinii, correspond to a fragment of apolipoprotein B, a lipoprotein that functions as a ligand for the low-density lipoprotein (LDL) receptor.
  • LDL low-density lipoprotein
  • NPN a lipophilic dye
  • NPN fluoresces faintly in aqueous solutions but fluoresces substantially more when it encounters lipidic environments such as bacterial membranes.
  • NPN can penetrate the bacterial outer membrane only if it is disrupted or compromised.
  • bacteria exposed to the most active EPs at their MIC were, in general, not effectively permeabilized (FIG.5d and FIG.54).
  • the latter is an intrinsically resistant bacterium associated with infections of the urinary tract, gastrointestinal tissue, 24-10510 (103241.007081) skin, and soft tissues and a cause of bacterial pneumonia, as well as a common opportunistic pathogen in cystic fibrosis patients 31 .
  • Most of the combinations tested resulted in synergistic or additive interactions, calculated by using the fractional inhibitory concentration index 32 (FICI, FIG.5e).
  • the peptides in these lists were ranked increasingly by the ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ and the ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇
  • Bacterial strains and growth conditions used in the experiments [00160] The following Gram-negative bacteria were used in the present study: Acinetobacter baumannii ATCC 19606, Escherichia coli ATCC 11775, Escherichia coli AIC221 (Escherichia coli MG1655 phnE_2::FRT), Escherichia coli AIC222 [Escherichia coli MG1655 pmrA53 phnE_2::FRT (colistin-resistant)], Klebsiella pneumoniae ATCC 13883, Pseudomonas aeruginosa PAO1, and Pseudomonas aeruginosa PA14.
  • Antibacterial assays [00161] The in vitro antimicrobial activity of the peptides was assessed by subjecting them to the broth microdilution assay 31 . Minimum inhibitory concentration (MIC) values of the peptides were determined with an initial inoculum of 2 ⁇ 10 6 cells mL -1 in LB in microtiter 96-well flat bottom transparent plates.
  • Aqueous solutions of the 24-10510 (103241.007081) peptides were added to the plate at concentrations ranging from 1 to 64 ⁇ mol L -1 .
  • the lowest concentration of peptide that inhibited 100% of the visible growth of bacteria was established as the MIC value in an experiment of 20 h of exposure at 37 °C.
  • the optical density of the plates was measured at 600 nm using a spectrophotometer. All assays were done as three biological replicates.
  • Outer membrane permeabilization assays [00162] The membrane permeability of the peptides was determined by using the 1-(N-phenylamino)naphthalene (NPN) uptake assay 12 .
  • NPN is a hydrophobic fluorescent dye that does not readily permeate the bacterial outer membrane. However, when the membrane integrity is compromised, NPN can enter the cell and bind to the bacterial membrane lipids. This causes the dye to exhibit a strong fluorescence.
  • the cells were then centrifuged using the same conditions described for the NPN uptake assays, washed twice with washing buffer containing 20 mmol L -1 glucose and 5 mmol L -1 HEPES (pH 24-10510 (103241.007081) 7.2).
  • One hundred ⁇ L of bacterial solution were then incubated for 15 min with 20 nmol L -1 of DiSC3-5 until the fluorescence reached a plateau, i.e., the dye was fully internalized into the bacterial membrane.
  • HEK293T Human embryonic kidney cells were obtained from the American Type Culture Collection (ATCC; CRL-3216TM).
  • the cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 1% penicillin and streptomycin (antibiotics) and 10% fetal bovine serum (FBS) and grown at 37 °C in a humidified atmosphere containing 5% CO2.
  • DMEM high-glucose Dulbecco’s modified Eagle’s medium
  • FBS fetal bovine serum
  • MTT 24-10510 (103241.007081) assay tetrazolium reduction assay 49 .
  • the MTT reagent was dissolved at 0.5 mg mL -1 in medium without phenol red and was used to replace cell culture supernatants containing the peptides (100 ⁇ L per well), and the samples were incubated for 4 h at 37 °C in a humidified atmosphere containing 5% CO 2 yielding the insoluble formazan salt.
  • the mobile phases used were A (100% water with 0.1%, v/v, formic acid) and B (100% acetonitrile with 0.1%, v/v, formic acid), Fisher optima grades. Measurements were made by ionization ESI +/- simultaneous over m/z 100-2,000 Da. The percentage of remaining peptide was calculated by integrating the area under the curve related to the peptide at time point zero. Experiments were performed in three independent replicates. Time A Flow rate (mL B (%) (min) (%) min -1 ) 0 95 5 0.5 0.5 95 5 0.5 0.5 (linear 2.5 5 95 gradient) 24-10510 (103241.007081) 3 5 95 0.5 3.25 0.5 Skin abscess infection mouse model [00167] A.
  • CFU colony-forming units
  • Thigh infection mouse model Six-week-old female CD-1 mice from Charles River (stock number 18679700-022) were rendered neutropenic by two doses of cyclophosphamide (150 mg Kg -1 ) applied intraperitoneally with an interval of 72 h.
  • mice were injected intramuscularly in their right thigh with a bacterial load of 10 6 CFU mL -1 of A. baumannii ATCC 19606 cells.
  • the bacteria had been grown in LB broth, washed twice with PBS (pH 7.4), and resuspended to the desired concentration.
  • PBS pH 7.4
  • peptides resuspended in water were 24-10510 (103241.007081) administered intraperitoneally.
  • mice Prior to each injection, mice were anesthetized with isoflurane and monitored for respiratory rate and pedal reflexes. Next, we monitored the establishment of the infection and euthanized the mice.
  • the infected area was excised two days and four days post-infection, homogenized using a bead beater for 20 min (25 Hz), and 10-fold serially diluted for CFU quantification in MacConkey agar plates.
  • the experiments were performed with 6 mice per group. All experiments were performed blindly, and no animal subjects were excluded from the analysis.
  • the thigh infection mouse model was approved by the University Laboratory Animal Resources (ULAR) from the University of Pennsylvania (Protocol 807055). Statistical significance was determined using one-way ANOVA followed by Dunnett’s test in a log 10 -transformed data to mitigate the effect of outliers; p values are presented for each group, with all groups being compared to the untreated control group.
  • Hyperparameter ranges searched for elastic net Elastic Net hyperparameter alpha Hyperparameter range ⁇ 1.0, 0.5, 0.1, 0.05, 0.01, 0.005, 0.001, 0.0005, 0.0001 ⁇ Table S3.
  • Linear support vector regression hyperparameter C Hyperparameter range ⁇ 1, 10, 20, 30, ..., 1000 ⁇ Table S4.
  • Table S6 Pearson correlation of various ML models on CV set.
  • the table shows the average species-specific prediction performance of 5-fold CV in terms of Pearson correlation for various ML models on the CV set.
  • RF random forest
  • GBDT gradient boosting decision tree
  • ExtraTree extra-tree regressor
  • ElasticNet elastic net
  • LinearSVR linear support vector regression.
  • Table S7 Spearman correlation of various ML models on CV set.
  • the table shows the average species-specific prediction performance of 5-fold CV in terms of Spearman correlation for various ML models on the CV set.
  • RF random forest
  • GBDT gradient boosting decision tree
  • ExtraTree extra-tree regressor
  • ElasticNet elastic net
  • LinearSVR linear support vector regression.
  • Table S8 R-squared of single APEX variants on CV set.
  • the table shows the average species-specific prediction performance of 5-fold CV in terms of R-squared for various APEX variants, including the original APEX (i.e., single APEX), APEX without multitask constraint, APEX without using public AMP data during training, and APEX without multitask constraint and public AMP data during training.
  • Table S11 Hyperparameters of top eight APEX ranked by R-squared on CV set. Top eight APEX models number of RNN layers n m ⁇ _l2 ⁇ _(multitask_constraint) ⁇ _BCE 1 3 128 2048 1.00E-05 0.1 1 2 3 256 2048 1.00E-06 0.1 1 3 2 128 512 1.00E-05 0.01 1 4 3 128 512 1.00E-05 0.001 1 5 2 128 2048 1.00E-06 0 1 6 3 256 512 1.00E-06 0 1 7 2 128 2048 1.00E-05 0.01 1 8 2 256 2048 1.00E-06 0.1 1
  • Table S14 Spearman correlation of various ML models on independent set.
  • the table shows the species- specific prediction performance in terms of Spearman correlation for various ML models that were trained on the CV set and evaluated on the independent set.
  • RF random forest
  • GBDT gradient boosting decision tree
  • ExtraTree extra-tree regressor
  • ElasticNet elastic net
  • LinearSVR linear support vector regression. Strain Ensemble Ensemble Single APEX v2 APEX v1 APEX RF GBDT ExtraTree ElasticNet LinearSVR E.
  • aeruginosa PAO1 0.710869 0.686719 0.579761 0.731061 0.679557 0.699833 0.569554 0.143789 0.582418
  • P. aeruginosa PA14 0.734694 0.73291 0.671569 0.749065 0.68225 0.749057 0.725144 0.254412 0.712036
  • S. aureus ATCC12600 0.126253 0.186218 0.199264 0.076771 0.020007 0.07906 -0.01304 -0.26126 0.42482
  • E. coli AIC221 0.403527 0.404244 0.389027 0.521815 0.413622 0.483885 0.398567 0.200215 0.311953
  • aeruginosa PAO1 0.719366 0.718463 0.622332 0.673903 0.575407 0.648483 0.626982 0.03871 0.615923 P. aeruginosa 0.76658 0.760304 0.657554 0.657068 0.683707 0.662736 0.688632 0.225784 0.721803 0.025934 -0.20224 0.4654 0.409659 0.137742 0.31409 0.55179 0.17471 0.477801 0.440967 0.066277 0.293077 0.513264 0.458603 0.455221 0.462197 0.140025 0.480142
  • Cytotoxic activity of AEPs and MEPs The cytotoxic activity was expressed in terms of CC 50 values ( ⁇ mol L -1 ), i.e., cytotoxic concentration values needed to damage 50% of the HEK293T cells present in each condition. The values were estimated by non-linear regressions based on the screen of all active AEPs and MEPs at concentrations from 8 to 128 ⁇ mol L -1 , to ensure coverage of all tested antimicrobial activity concentrations. The experiments were done in three independent biological replicates with two technical replicates within each biological replicate.
  • the therapeutic index (TI) was calculated to show the margin of safety obtained by comparing the lowest MIC values ( ⁇ mol L -1 ) obtained in the antimicrobial activity assays to the CC50 values of each active AEP or MEP.
  • Peptide CC50 ( ⁇ mol L -1 ) MIC ( ⁇ mol L -1 ) TI
  • Peptide CC50 ( ⁇ mol L -1 ) MIC ( ⁇ mol L -1 ) TI Equusin-1 >128 1 >128 Megalocerin-1 >128 8 >16 Hesperelin-1 >128 2 >64 Pinguinusin-1 >128 4 >32 Elephasin-1 >128 4 >32
  • Antimicrobial potency of cationic antimicrobial peptides can be predicted from their amino acid composition: Application to the detection of “cryptic” antimicrobial peptides. J Theor Biol 419, 254–265 (2017). 19. Zhao, M., Lee, W.-P., Garrison, E. P. & Marth, G. T. SSW Library: An SIMD Smith- Waterman C/C++ Library for Use in Genomic Applications. PLoS One 8, e82138 (2013). 20. McInnes, L., Healy, J., Saul, N. & cateberger, L. UMAP: Uniform Manifold Approximation and Projection. J Open Source Softw 3, 861 (2016). 21. Torres, M. D.
  • Drug combinations a strategy to extend the life of antibiotics in the 21st century. Nat Rev Microbiol 17, 141–155 (2019). 33. Lázár, V., Snitser, O., Barkan, D. & Kishony, R. Antibiotic combinations reduce Staphylococcus aureus clearance. Nature 610, 540–546 (2022). 34. Nim, S. et al. Disrupting the ⁇ -synuclein-ESCRT interaction with a peptide inhibitor mitigates neurodegeneration in preclinical models of Parkinson’s disease. Nat Commun 14, 2150 (2023). 35. Silva, O. N. et al.

Landscapes

  • Health & Medical Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Organic Chemistry (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Engineering & Computer Science (AREA)
  • Medicinal Chemistry (AREA)
  • Medical Informatics (AREA)
  • Biochemistry (AREA)
  • Molecular Biology (AREA)
  • Proteomics, Peptides & Aminoacids (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Genetics & Genomics (AREA)
  • Public Health (AREA)
  • Epidemiology (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Veterinary Medicine (AREA)
  • Gastroenterology & Hepatology (AREA)
  • General Chemical & Material Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Oncology (AREA)
  • Artificial Intelligence (AREA)
  • Bioethics (AREA)
  • Animal Behavior & Ethology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Computation (AREA)
  • Communicable Diseases (AREA)
  • Software Systems (AREA)
  • Pharmacology & Pharmacy (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Biotechnology (AREA)
  • Evolutionary Biology (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Theoretical Computer Science (AREA)
  • Medicines That Contain Protein Lipid Enzymes And Other Medicines (AREA)
  • Peptides Or Proteins (AREA)

Abstract

L'invention concerne des procédés d'identification de peptides antimicrobiens provenant de protéomes éteints à l'aide d'une approche d'apprentissage profond multitâche. L'invention concerne également des peptides antimicrobiens, et des procédés de traitement d'une infection microbienne comprenant la mise en contact de l'infection avec les présents peptides antimicrobiens. L'invention concerne par ailleurs des méthodes de traitement d'une infection microbienne comprenant l'administration à un sujet qui en a besoin d'une quantité pharmaceutiquement efficace d'un peptide antimicrobien selon la présente divulgation.
PCT/US2024/045821 2023-09-07 2024-09-09 Désextinction moléculaire d'antibiotiques activés par apprentissage profond Pending WO2025054593A2 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US202363581122P 2023-09-07 2023-09-07
US63/581,122 2023-09-07

Publications (2)

Publication Number Publication Date
WO2025054593A2 true WO2025054593A2 (fr) 2025-03-13
WO2025054593A3 WO2025054593A3 (fr) 2025-07-31

Family

ID=94924305

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2024/045821 Pending WO2025054593A2 (fr) 2023-09-07 2024-09-09 Désextinction moléculaire d'antibiotiques activés par apprentissage profond

Country Status (1)

Country Link
WO (1) WO2025054593A2 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN121148732A (zh) * 2025-09-28 2025-12-16 首都医科大学附属北京潞河医院 一种慢性静脉功能不全的分型分级系统

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112368376A (zh) * 2018-03-29 2021-02-12 康特拉费克特公司 溶素-抗微生物肽(amp)多肽构建体、溶素、其分离的编码多核苷酸及其用途
WO2022217285A1 (fr) * 2021-04-09 2022-10-13 The Trustees Of The University Of Pennsylvania Antibiotiques cachés présents dans le protéome humain

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN121148732A (zh) * 2025-09-28 2025-12-16 首都医科大学附属北京潞河医院 一种慢性静脉功能不全的分型分级系统

Also Published As

Publication number Publication date
WO2025054593A3 (fr) 2025-07-31

Similar Documents

Publication Publication Date Title
Wan et al. Deep-learning-enabled antibiotic discovery through molecular de-extinction
Torres et al. Mining human microbiomes reveals an untapped source of peptide antibiotics
Santos-Júnior et al. Discovery of antimicrobial peptides in the global microbiome with machine learning
Rončević et al. Antimicrobial peptides as anti-infective agents in pre-post-antibiotic era?
Torres et al. Mining for encrypted peptide antibiotics in the human proteome
Lakshmaiah Narayana et al. Two distinct amphipathic peptide antibiotics with systemic efficacy
Jiang et al. “Specificity determinants” improve therapeutic indices of two antimicrobial peptides piscidin 1 and dermaseptin S4 against the Gram-negative pathogens Acinetobacter baumannii and Pseudomonas aeruginosa
Wang Database-guided discovery of potent peptides to combat HIV-1 or superbugs
Czyzewski et al. In vivo, in vitro, and in silico characterization of peptoids as antimicrobial agents
Lee et al. Structure–activity relationships of de novo designed cyclic antimicrobial peptides based on gramicidin S
Hansen et al. Antimicrobial activity of small synthetic peptides based on the marine peptide turgencin a: Prediction of antimicrobial peptide sequences in a natural peptide and strategy for optimization of potency
Wang et al. Mirror images of antimicrobial peptides provide reflections on their functions and amyloidogenic properties
Panteleev et al. Structure elucidation and functional studies of a novel β-hairpin antimicrobial peptide from the marine polychaeta Capitella teleta
US20240194290A1 (en) Hidden antibiotics in the human proteome
Hansen et al. Structure of biofilm-forming functional amyloid PSMα1 from Staphylococcus aureus
Wan et al. Molecular de-extinction of antibiotics enabled by deep learning
Torres et al. Generative latent diffusion language modeling yields anti-infective synthetic peptides
Richter et al. Associating biological activity and predicted structure of antimicrobial peptides from amphibians and insects
Hervé et al. Three-dimensional NMR structure of hen egg gallin (chicken ovodefensin) reveals a new variation of the β-defensin fold
WO2025054593A2 (fr) Désextinction moléculaire d'antibiotiques activés par apprentissage profond
Torres et al. A generative artificial intelligence approach for antibiotic optimization
Ludwig et al. Antimicrobial activity and 70S ribosome binding of apidaecin-derived Api805 with increased bacterial uptake rate
Howan et al. Enhanced antibacterial activity of substituted derivatives of NCR169C peptide
Dalla Torre et al. Peptaibol analogs show potent antibacterial activity against multidrug resistant opportunistic pathogens
Torres et al. Human gut metagenomic mining reveals an untapped source of peptide antibiotics

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 24863774

Country of ref document: EP

Kind code of ref document: A2