Phylogeography of Bacterial Pathogens

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-27

Quick Answer

Phylogeography links bacterial genomic variation to geographic origin, letting researchers trace transmission routes and identify outbreak sources from whole-genome sequence data.
The practical workflow requires curated genome collections, quality-filtered sequence alignments, a substitution model matched to your data, and a phylogeographic model that accounts for sampling bias.
A key limitation is that sparse or uneven geographic sampling can produce false migration patterns, so results require validation against epidemiological metadata and independent surveillance data.

At a Glance

Analysis Stage	Primary Decision	Common Data Input	Typical Output
Data assembly	Select genomes with complete geographic metadata	FASTA files, GenBank records	Curated sequence set with location tags
Quality control	Filter low-coverage or contaminated assemblies	Raw reads or assembled contigs	High-confidence SNP alignment
Model selection	Choose substitution model based on dataset size	Alignment file, IQ-TREE or BEAST input	Maximum likelihood tree or Bayesian posterior
Phylogeographic inference	Balance discrete vs continuous models	Tree file, location annotations	Ancestral state reconstruction with support values
Validation	Compare against epidemiological records	Case counts, travel histories	Concordance assessment

What Phylogeography Reveals About Bacterial Pathogens

Phylogeography connects the evolutionary relationships among bacterial isolates to their geographic origins. The approach uses genomic sequences to build phylogenetic trees and then maps the inferred ancestral states onto spatial locations. This allows researchers to ask whether isolates from the same region cluster together, whether a pathogen moved between countries, and which geographic routes best explain the observed distribution of genetic diversity.

The core principle is that bacterial genomes accumulate mutations over time. When isolates share recent common ancestry, they carry similar mutations. If those isolates also share a geographic origin, the phylogenetic tree will show geographic structure. When isolates from different regions appear on the same branch, the tree suggests a transmission event between those regions.

This approach has direct applications in outbreak investigation. During a foodborne outbreak, public health laboratories sequence isolates from patients and from suspected food sources. Phylogeographic analysis can show whether the patient isolates match the food isolates and whether the outbreak strain originated from a specific production facility or region. The same logic applies to hospital-acquired infections, where isolates from different wards or different hospitals can be compared to identify transmission routes.

The World Health Organization recognizes genomic surveillance as a core component of the global response to antimicrobial resistance. Tracking resistant pathogens across geographic boundaries requires the same phylogeographic tools used for outbreak investigation. When a resistant clone appears in multiple countries, the tree can show whether those appearances are independent acquisitions or a single clone spreading across borders.

Core Principles of Bacterial Phylogeography

The Relationship Between Mutation and Geography

Bacterial evolution proceeds through mutation, recombination, and selection. For phylogeographic inference, the most useful mutations are single nucleotide polymorphisms that accumulate at a relatively steady rate. These mutations create a molecular clock that allows researchers to estimate when two isolates last shared a common ancestor.

The geographic signal comes from the association between the tree and the sampling locations. If a bacterial population has been established in a region for a long time, the isolates from that region will show deep branches that reflect local diversification. If a pathogen has recently invaded a region, the isolates will cluster together on a short branch that connects to an ancestor from elsewhere.

This logic assumes that the geographic location of an isolate reflects the location of its ancestor. That assumption fails when a patient has traveled, when a food product has been shipped, or when a sample has been contaminated in the laboratory. These sources of error must be considered when interpreting the tree.

Discrete Versus Continuous Models

Phylogeographic models fall into two broad categories. Discrete models treat geography as a set of categories, such as country, state, or city. The model estimates the probability that an ancestor in one location gave rise to a descendant in another location. This approach is simple and works well when the sampling locations are clearly defined.

Continuous models treat geography as a two-dimensional space. The model estimates the geographic coordinates of ancestral nodes and the rate of movement across the landscape. This approach is more powerful when the data include precise coordinates, but it is also more sensitive to sampling bias.

The choice between discrete and continuous models depends on the research question. A discrete model is appropriate for comparing countries or states. A continuous model is appropriate for tracking the spread of a pathogen across a landscape or between cities.

The Role of Sampling Bias

Sampling bias is the most serious threat to phylogeographic inference. If a study includes many isolates from one region and few from another, the tree will show more diversity in the well-sampled region. This can create the appearance of a migration event when the true pattern is simply a difference in sampling effort.

The standard approach to this problem is to include a sampling bias correction in the model. The correction accounts for the number of isolates from each location and adjusts the migration rates accordingly. Researchers should also consider whether the sampling strategy reflects the true distribution of the pathogen or the convenience of the laboratory.

The World Health Organization notes that surveillance data are only as good as the systems that produce them. The same principle applies to phylogeographic data. A tree built from a biased sample will produce a biased picture of the geographic spread.

Data Preparation for Phylogeographic Analysis

Genome Collection and Metadata

The first step is to assemble a collection of genomes with complete metadata. Each genome must have a unique identifier, a collection date, and a geographic location. The location should be as specific as possible, ideally at the city or district level. The collection date is essential for the molecular clock analysis.

The American Type Culture Collection provides authenticated bacterial strains that can serve as reference material. When building a phylogeographic dataset, researchers should include reference genomes from the ATCC or other culture collections to anchor the tree and to verify the quality of the sequencing data.

The metadata should be stored in a structured format that can be read by the analysis software. Common formats include CSV, TSV, and JSON. The metadata file must be checked for errors before the analysis begins. A single incorrect location tag can produce a false migration event.

Quality Control of Genomic Data

The quality of the input genomes determines the quality of the tree. Low coverage genomes produce ambiguous base calls that can be mistaken for mutations. Contaminated genomes can produce sequences from two different organisms, which will create a false branch in the tree.

The quality control step should include the following checks:

Coverage depth: genomes with low coverage should be excluded or re-sequenced
Contamination: the presence of multiple species or strains should be detected and removed
Assembly quality: the number of contigs and the N50 value should be recorded
Read mapping: the reads should be mapped to a reference genome to identify variants

The Clinical and Laboratory Standards Institute provides standards for the quality of microbiological testing. While these standards are designed for antimicrobial susceptibility testing, the same principle of quality control applies to genomic data. A laboratory that produces unreliable susceptibility results will also produce unreliable genomic data.

Alignment and Variant Calling

The next step is to align the genomes to a reference and call variants. The reference genome should be a high-quality, complete genome of the same species. The alignment should be done with a tool that is designed for bacterial genomes, such as a whole-genome aligner.

The variant calling step produces a list of single nucleotide polymorphisms that are used to build the tree. The SNPs must be filtered to remove those that are in repetitive regions or that are likely to be the result of recombination. Recombination can create false phylogenetic signals because it transfers large blocks of DNA between strains.

The filtered SNP list is then used to build the alignment that will be the input to the phylogenetic analysis. The alignment should include only the variable sites, or the entire genome with the variable sites marked. The choice depends on the analysis software.

Model Selection for Phylogeographic Inference

Substitution Models

The substitution model describes how the nucleotides change over time. The simplest model is the Jukes-Cantor model, which assumes that all nucleotide changes are equally likely. More complex models allow for different rates of transition and transversion, and for different base frequencies.

The model selection should be based on the data. The model that best fits the data can be identified using a model selection tool that compares the likelihood of different models. The selected model should be used for the tree building and the phylogeographic inference.

The choice of substitution model has a direct effect on the estimated branch lengths. A model that is too simple will underestimate the branch lengths, which will affect the molecular clock and the migration rates. A model that is too complex will overfit the data and produce unstable estimates.

Molecular Clock Models

The molecular clock describes the rate at which mutations accumulate. A strict clock assumes that all branches have the same rate. A relaxed clock allows the rate to vary among branches. The choice of clock model depends on the data and the research question.

The strict clock is appropriate when the pathogen has a stable mutation rate and the sampling time is short. The relaxed clock is appropriate when the mutation rate varies among lineages or when the sampling period is long. The relaxed clock requires more data to estimate the rate variation.

The clock rate is estimated from the data using the collection dates. The rate is expressed in substitutions per site per year. The rate can be compared to published rates for the same species to verify the estimate.

Discrete Phylogeographic Models

The discrete phylogeographic model treats the geographic locations as discrete states. The model estimates the rate of migration between each pair of locations. The migration rates are then used to reconstruct the ancestral locations of the internal nodes of the tree.

The model can be implemented in a Bayesian framework, which produces a posterior distribution of the migration rates and the ancestral locations. The Bayesian approach requires a prior distribution for the migration rates. The prior should be chosen based on the expected rate of migration for the pathogen.

The output of the discrete model is a set of ancestral location estimates for each node. The estimates are accompanied by a posterior probability, which indicates the confidence in the estimate. A posterior probability of 0.9 or higher is considered strong support.

Continuous Phylogeographic Models

The continuous phylogeographic model treats the geographic locations as coordinates in a two-dimensional space. The model estimates the location of the ancestor of each node and the rate of movement across the space.

The continuous model is more complex than the discrete model. It requires the geographic coordinates of each isolate and a model of how the pathogen moves across the landscape. The model can be implemented in a Bayesian framework, but the computational cost is higher.

The output of the continuous model is a set of estimated ancestral locations. The locations can be plotted on a map to show the spread of the pathogen over time. The map can be used to identify the origin of the outbreak and the routes of transmission.

Practical Workflow for a Phylogeographic Analysis

Step 1: Define the Research Question

The first step is to define the research question. The question should be specific and answerable with the available data. Examples include:

Did the outbreak strain originate from a single source?
Is the resistant clone spreading between countries?
What is the geographic origin of the local outbreak?

The research question determines the type of data that is needed and the model that is used. A question about the origin of an outbreak requires a discrete model with the locations of the isolates. A question about the rate of spread requires a continuous model with the coordinates of the isolates.

Step 2: Assemble the Genome Collection

The genome collection should include all available isolates that are relevant to the research question. The collection should include isolates from the outbreak and isolates from the surrounding region. The collection should also include reference genomes from the same species.

The metadata for each isolate should be verified. The collection date should be accurate. The geographic location should be specific. The source of the isolate should be recorded, such as the patient, the food, or the environment.

Step 3: Perform Quality Control

The quality control step should be performed before the analysis. The genomes should be checked for coverage, contamination, and assembly quality. The genomes that fail the quality control should be excluded or re-sequenced.

The quality control should be documented. The documentation should include the number of genomes that were excluded and the reason for the exclusion. The documentation should be included in the final report.

Step 4: Build the Alignment

The alignment should be built from the filtered SNP list. The alignment should be checked for errors. The alignment should be used as the input to the phylogenetic analysis.

The alignment should be stored in a standard format, such as FASTA or PHYLIP. The alignment should be accompanied by a file that describes the metadata for each sequence.

Step 5: Select the Model

The model should be selected based on the data. The substitution model should be selected using a statistical comparison. The clock model should be selected based on the expected mutation rate. The phylogeographic model should be selected based on the research question.

The model selection should be documented. The documentation should include the model that was selected and the reason for the selection.

Step 6: Run the Analysis

The analysis should be run with the selected model. The analysis should be run with a sufficient number of iterations to ensure convergence. The convergence should be checked using the trace of the parameters.

The analysis should be run multiple times to ensure that the results are reproducible. The results should be compared across the runs. The results should be consistent across the runs.

Step 7: Interpret the Results

The results should be interpreted in the context of the research question. The ancestral locations should be examined. The migration rates should be examined. The posterior probabilities should be examined.

The results should be compared to the epidemiological data. The results should be consistent with the epidemiological data. If the results are not consistent, the data or the model should be re-examined.

Records and Measurements

Data Management

The data management plan should include the storage of the raw reads, the assembled genomes, the alignment, and the analysis results. The data should be stored in a secure location. The data should be backed up.

The data should be organized in a way that is easy to retrieve. The data should be labeled with the date and the version. The data should be stored in a format that is readable by the analysis software.

Quality Metrics

The quality metrics should be recorded for each genome. The metrics should include the coverage depth, the number of contigs, the N50 score, and the number of SNPs. The metrics should be recorded in a table.

The quality metrics should be used to filter the genomes. The genomes that do not meet the quality threshold should be excluded. The quality threshold should be defined before the analysis.

Analysis Log

The analysis log should record the steps of the analysis. The log should include the software version, the model parameters, and the date of the analysis. The log should be stored with the analysis results.

The analysis log should be used to reproduce the analysis. The log should be complete enough to allow another researcher to reproduce the analysis.

Common Failure Patterns

Sampling Bias

The most common failure is sampling bias. The tree will show a migration event when the true pattern is a difference in sampling. The bias can be reduced by including a sampling bias correction in the model.

The bias can be detected by comparing the number of isolates per location to the number of branches per location. If the number of isolates is high and the number of branches is low, the sampling is biased.

Recombination

Recombination can create false mutations in the alignment. The false mutations can produce a false branch in the tree. The recombination can be detected by examining the SNP distribution across the genome.

The recombination can be removed by filtering the SNPs that are in the recombinant regions. The filtering should be performed before the tree is built.

The Clock

The clock can be misspecified. The clock can be too fast or too slow. The misspecified clock can produce an incorrect time of the ancestor.

The clock can be checked by comparing the estimated rate to the published rate. The clock can be checked by comparing the estimated time to the known time of the outbreak.

The Sampling

The sampling can be biased. The sampling can be biased toward a specific region or a specific source. The biased sampling can produce a misleading tree.

The sampling can be improved by including isolates from a wider range of locations and sources. The sampling can be improved by including isolates from the environment and the animal reservoir.

Limitations of Phylogeographic Inference

The Resolution of the Data

The resolution of the data is limited by the number of isolates and the number of SNPs. The resolution is limited by the mutation rate of the pathogen. The resolution is limited by the sampling period.

The resolution can be improved by increasing the number of isolates and the number of SNPs. The resolution can be improved by including isolates from a longer time period.

The Assumptions of the Model

The model assumes that the mutation rate is constant. The model assumes that the migration rate is constant. The model assumes that the sampling is unbiased.

The assumptions are not always met. The mutation rate can vary among lineages. The migration rate can vary over time. The sampling can be biased.

The Interpretation of the Results

The results of the phylogeographic analysis are a statistical estimate. The results are not a detailed explanation. The results should be interpreted in the context of the epidemiological data.

The results should be used to generate hypotheses. The hypotheses should be tested with additional data. The hypotheses should be tested with additional analyses.

Biosafety and Laboratory Practice

Risk Assessment

The phylogeographic analysis requires the handling of bacterial isolates. The isolates should be handled in a biosafety cabinet. The isolates should be handled according to the risk assessment.

The risk assessment should be performed before the analysis. The risk assessment should consider the species of the bacteria, the source of the isolate, and the procedure that is being performed. The risk assessment should be documented.

The CDC and NIH provide guidance on the biosafety of microbiological laboratories. The guidance includes the recommended biosafety level for different species of bacteria. The guidance includes the recommended practices for the handling of the bacteria.

Decontamination

The decontamination should be performed after the analysis. The decontamination should be performed on the work surface and the equipment. The decontamination should be performed with a disinfectant that is effective against the bacteria.

The decontamination should be documented. The documentation should include the disinfectant that was used and the time of the decontamination.

The Use of Authenticated Strains

The use of authenticated strains is important for the quality of the analysis. The authenticated strains should be obtained from a culture collection, such as the ATCC. The authenticated strains should be used as a reference for the analysis.

The authenticated strains should be stored according to the instructions of the culture collection. The authenticated strains should be checked for the viability and the purity.

The Role of Phylogeography in Public Health

Outbreak Investigation

The phylogeographic analysis is used in the outbreak investigation. The analysis is used to identify the source of the outbreak. The analysis is used to identify the route of the transmission.

The analysis is used to support the epidemiological investigation. The analysis is used to confirm the epidemiological findings. The analysis is used to provide the evidence for the control measures.

The Surveillance of Antimicrobial Resistance

The phylogeographic analysis is used in the surveillance of antimicrobial resistance. The analysis is used to track the spread of the resistant bacteria. The analysis is used to identify the origin of the resistant bacteria.

The World Health Organization has identified antimicrobial resistance as a global health threat. The surveillance of the resistant bacteria is a key component of the response. The phylogeographic analysis is a tool for the surveillance.

The Food Safety

The phylogeographic analysis is used in the food safety. The analysis is used to identify the source of the foodborne outbreak. The analysis is used to identify the route of the contamination.

The FDA provides the Bacteriological Analytical Manual for the detection of the foodborne bacteria. The BAM includes the methods for the isolation and the identification of the bacteria. The phylogeographic analysis is used to support the BAM methods.

The Limitations of the Evidence

The phylogeographic analysis is a statistical estimate. The analysis is based on the data that is available. The analysis is based on the model that is selected.

The analysis is limited by the quality of the data. The analysis is limited by the assumptions of the model. The analysis is limited by the interpretation of the results.

The analysis should be used in the context of the epidemiological data. The analysis should be used to support the epidemiological findings. The analysis should not be used as the only evidence.

The Professional Escalation Criteria

The phylogeographic analysis should be performed by a qualified researcher. The researcher should have the training in the bioinformatics and the phylogenetics. The researcher should have the experience with the bacterial genomics.

The analysis should be reviewed by a second researcher. The review should include the data, the model, and the results. The review should be documented.

The analysis should be escalated to the public health authority if the results indicate a public health threat. The public health authority should be notified of the results. The public health authority should be provided with the data and the analysis.

The Future of Phylogeographic Analysis

The phylogeographic analysis is a rapidly evolving field. The new methods are being developed. The new methods are being applied to the new data.

The new methods include the machine learning and the artificial intelligence. The new methods are being used to improve the accuracy of the analysis. The new methods are being used to improve the speed of the analysis.

The new data includes the whole-genome sequencing data. The new data includes the metagenomic data. The new data includes the data from the environment.

The phylogeographic analysis is a powerful tool for the understanding of the spread of the bacterial pathogens. The analysis is a tool for the public health. The analysis is a tool for the food safety.

Decision Framework for Phylogeographic Study Design

The Study Design Triage

Before any sequence is downloaded or any alignment is built, a researcher must decide whether a phylogeographic analysis is the correct tool for the question at hand. This decision framework uses a structured triage that prevents wasted computation and misinterpretation. The triage asks three sequential questions that determine whether the analysis should proceed, be redesigned, or be abandoned in favor of a simpler approach.

The first question asks whether the research question requires geographic inference. If the question is only about evolutionary relationships among isolates, a standard phylogenetic tree is sufficient. Phylogeographic analysis adds computational complexity and model assumptions that are unnecessary when geography is not the focus. The second question asks whether the sampling design can support geographic inference. A dataset with isolates from one dominant location and a handful from elsewhere will not support a reliable migration analysis. The third question asks whether the epidemiological metadata is complete enough to validate the results. Without independent records of case locations, travel histories, or food distribution chains, the phylogeographic output cannot be checked against reality.

This triage produces one of three outcomes. The analysis proceeds with the proposed design, the analysis proceeds with a modified sampling or modeling strategy, or the analysis is deferred until additional data can be collected. The triage should be documented in the study protocol and revisited when new isolates are added to the dataset.

The Sampling Adequacy Assessment

The sampling adequacy assessment is a quantitative check that should be performed before the analysis begins. The assessment compares the number of isolates per location to the expected diversity of the pathogen in that location. A location with a long-established bacterial population will have more genetic diversity than a location with a recent introduction. If the sampling effort does not reflect this expected diversity, the phylogeographic model will produce biased migration estimates.

The assessment uses a simple ratio. For each location, divide the number of isolates by the number of distinct phylogenetic clusters observed in a preliminary tree. A ratio below one indicates that the location is undersampled. A ratio above five indicates that the location may be oversampled relative to its diversity. The preliminary tree can be built with a fast maximum likelihood method before the full Bayesian analysis is run.

The World Health Organization surveillance guidance emphasizes that the quality of surveillance data depends on the systems that produce it. The same principle applies to phylogeographic sampling. A dataset that is convenient to assemble, such as isolates from a single reference laboratory, will not represent the true geographic distribution of the pathogen. The sampling plan should be designed to match the research question, not the convenience of the laboratory.

The Model Selection Decision Tree

The model selection decision tree provides a structured path from the research question to the final model specification. The tree has four decision points that are evaluated in sequence.

The first decision point is the geographic scale. If the research question involves countries, states, or provinces, the discrete model is the default choice. If the research question involves cities, districts, or precise coordinates, the continuous model should be considered. The continuous model requires coordinates for each isolate and a model of movement across the landscape.

The second decision point is the sampling period. If the isolates were collected over a short period, such as a single outbreak, the strict clock is appropriate. If the isolates were collected over years or decades, the relaxed clock should be used. The relaxed clock allows the mutation rate to vary among branches, which is necessary when the pathogen has been evolving in different locations for different lengths of time.

The third decision point is the expected migration rate. If the pathogen is expected to move slowly, such as a soil-dwelling bacterium, the migration prior should be set to a low value. If the pathogen is expected to move quickly, such as a foodborne pathogen that travels through global supply chains, the migration prior should be set to a higher value. The prior should be based on published studies of the same or similar species.

The fourth decision point is the computational budget. The Bayesian phylogeographic analysis is computationally intensive. A large dataset with hundreds of isolates and a relaxed clock can take weeks to run. The researcher should estimate the computational time before the analysis begins and adjust the model complexity accordingly. A simpler model that runs to convergence is more useful than a complex model that does not converge.

The Sampling Bias Correction Protocol

The sampling bias correction protocol is a set of steps that should be applied when the dataset has an uneven distribution of isolates across locations. The protocol has three components.

The first component is the identification of the bias. The researcher should compare the number of isolates per location to the population size or the expected prevalence of the pathogen in that location. A location with a large population and a small number of isolates is undersampled. A location with a small population and a large number of isolates is over-sampled.

The second component is the correction. The correction can be applied in two ways. The first way is to include a sampling bias parameter in the model. The parameter accounts for the different sampling probabilities across locations. The second way is to subsample the over-sampled locations to match the sampling density of the under-sampled locations. The subsampling approach is simpler but discards data.

The third component is the sensitivity analysis. The analysis should be run with and without the correction to determine the effect of the bias on the results. If the results change substantially, the bias is having a strong effect and the corrected results should be reported. If the results are stable, the bias is having a minor effect and the uncorrected results may be acceptable.

The Validation and Escalation Protocol

The validation protocol is a structured comparison of the phylogeographic results against independent data. The protocol has three levels of validation.

The first level is the internal validation. The analysis is run multiple times with different random seeds to confirm that the results are reproducible. The posterior distributions of the migration rates and the ancestral locations should be similar across the runs. The convergence statistics should be checked to confirm that the Markov chain has reached a stable distribution.

The second level is the epidemiological validation. The inferred migration events are compared to the epidemiological records. The comparison includes the dates of the migration events, the locations of the migration events, and the sources of the isolates. A migration event that is not supported by the epidemiological records should be examined carefully. The discrepancy may indicate a data error, a model misspecification, or a true transmission event that was not detected by the epidemiological investigation.

The third level is the escalation protocol. The escalation protocol defines the conditions under which the results should be reported to a public health authority. The conditions include the identification of a new transmission route, the detection of a resistant clone in a new geographic region, or the identification of a food source that is linked to an outbreak. The World Health Organization has identified antimicrobial resistance as a global health threat that requires coordinated surveillance. The phylogeographic results that indicate the spread of a resistant clone across borders should be reported to the relevant public health authorities.

The escalation should be documented. The documentation should include the data, the analysis, the results, and the reason for the escalation. The documentation should be stored with the analysis records.

The Record System for Phylogeographic Analyses

The record system for phylogeographic analyses is a structured log that captures the decisions, the data, and the results of each analysis. The record system has five components.

The first component is the study protocol. The protocol describes the research question, the sampling plan, the model selection, and the validation plan. The protocol is written before the analysis begins and is updated when the analysis is revised.

The second component is the data log. The data log records the source of each genome, the collection date, the geographic location, and the quality metrics. The data log is used to track the data from the initial collection to the final analysis.

The third component is the analysis log. The analysis log records the software version, the model parameters, the date of the analysis, and the results. The analysis log is used to reproduce the analysis and to verify the results.

The fourth component is the validation log. The validation log records the results of the internal validation, the epidemiological validation, and the escalation protocol. The validation log is used to document the confidence in the results.

The fifth component is the decision log. The decision log records the decisions that were made during the analysis, including the model selection, the sampling bias correction, and the escalation. The decision log is used to document the rationale for the analysis choices.

The record system should be maintained in a secure location. The records should be backed up. The records should be accessible to the researchers who are involved in the analysis.

The Common Failure Patterns in Study Design

The study design failures are distinct from the analysis failures that are described in the existing article. The study design failures occur before the analysis begins and are often more costly to correct.

The first failure is the mismatch between the research question and the data. The research question asks about the geographic spread of a pathogen, but the data includes isolates from a single location. The analysis will produce a tree with no geographic signal, and the researcher will be unable to answer the question.

The second failure is the mismatch between the research question and the model. The research question asks about the rate of spread across a landscape, but the researcher uses a discrete model with countries as the locations. The discrete model cannot estimate the rate of spread across the landscape, and the researcher will need to redesign the analysis.

The third failure is the lack of a validation plan. The researcher builds the tree and infers the ancestral locations, but does not compare the results to the epidemiological data. The results may be misleading, and the researcher will not know.

The fourth failure is the lack of a computational budget. The researcher selects a complex model that requires a long computational time, but the analysis is not completed before the deadline. The researcher must simplify the model or extend the deadline.

The fifth failure is the lack of documentation. The researcher performs the analysis but does not record the decisions and the parameters. The analysis cannot be reproduced, and the results cannot be verified.

The Professional Escalation Criteria for Study Design

The escalation criteria for study design are distinct from the escalation criteria for the analysis results. The design escalation occurs when the researcher identifies a problem with the study design that requires the input of a more experienced researcher or a public health authority.

The first escalation criterion is the detection of a sampling bias that cannot be corrected. The researcher has identified a location that is severely under-sampled, and the correction is not possible with the available data. The escalation should be made to a researcher who has experience with the pathogen or the geographic region.

The second escalation criterion is the detection of a model misspecification. The researcher has identified a model that does not fit the data, and the alternative models are not available in the software. The escalation should be made to a researcher who has experience with the phylogeographic methods.

The third escalation criterion is the detection of a public health threat. The analysis has identified a new transmission route or a new geographic region that is affected by the pathogen. The escalation should be made to the public health authority, as described in the validation protocol.

The escalation should be documented. The documentation should include the reason for the escalation, the data and the analysis that support the escalation, and the response of the authority. The documentation should be stored in the record system.

The Professional Development and Training Context

The phylogeographic analysis requires a combination of skills that are not typically taught in a single course. The researcher needs the training in the bacterial genomics, the phylogenetics, and the geographic information systems. The researcher also needs the training in the epidemiological methods and the public health practice.

The American Society for Microbiology provides professional education and resources for the microbiology workforce. The ASM resources include the training in the laboratory practice, the quality control, and the professional development. The researcher should use the ASM resources to build the skills that are needed for the phylogeographic analysis.

The researcher should also seek the training in the specific software that is used for the phylogeographic analysis. The software has a learning curve, and the researcher should practice with the example datasets before the analysis of the real data. The practice should include the model selection, the analysis, and the interpretation of the results.

The researcher should also seek the mentorship from a researcher who has experience with the phylogeographic analysis. The mentorship can provide the guidance on the study design, the model selection, and the interpretation of the results. The mentorship can also provide the review of the analysis and the validation of the results.

The professional development should be documented. The documentation should include the training that was completed, the software that was learned, and the analyses that were performed. The documentation should be stored in the record system.

Frequently Asked Questions

What is the difference between a phylogenetic tree and a phylogeographic tree?

A phylogenetic tree shows the evolutionary relationships among isolates. A phylogeographic tree adds the geographic location of each isolate to the tree. The phylogeographic tree is used to infer the geographic spread of the pathogen.

What is the minimum number of isolates needed for a phylogeographic analysis?

The minimum number of isolates depends on the research question and the model. A simple analysis can be performed with a small number of isolates. A complex analysis requires a larger number of isolates. The number of isolates should be sufficient to provide the statistical power.

How do I choose between a discrete and a continuous phylogeographic model?

The choice depends on the research question and the data. A discrete model is appropriate when the locations are discrete, such as countries or states. A continuous model is appropriate when the locations are continuous, such as coordinates on a map.

What is the role of the molecular clock in phylogeographic analysis?

The molecular clock is used to estimate the time of the ancestor. The clock is used to estimate the rate of the mutation. The clock is used to estimate the time of the migration.

How do I account for sampling bias in my analysis?

The sampling bias can be accounted for by including a sampling bias correction in the model. The correction accounts for the number of isolates per location. The correction accounts for the number of isolates per source.

What is the difference between a Bayesian and a maximum likelihood approach?

The Bayesian approach produces a probability distribution of the parameters. The maximum likelihood approach produces a single estimate of the parameters. The Bayesian approach is more flexible and can incorporate prior information.

How do I validate the results of a phylogeographic analysis?

The results should be validated by comparing them to the epidemiological data. The results should be validated by comparing them to the independent data. The results should be validated by comparing them to the results of the other analyses.

What are the common mistakes in a phylogeographic analysis?

The common mistakes include the sampling bias, the recombination, and the misspecification of the model. The common mistakes include the misinterpretation of the results. The common mistakes should be avoided by the careful design of the analysis.

Using the Evidence

Source	Best use in this topic	Important limitation
Biosafety in Microbiological and Biomedical Laboratories	official guidance	Check the linked page for current local requirements
Microbiology Resources	official guidance	Check the linked page for current local requirements
Antimicrobial Resistance	official guidance	Check the linked page for current local requirements

Related Bacteriology Guides

References and Further Reading

Biosafety in Microbiological and Biomedical Laboratories. CDC and NIH.
Microbiology Resources. American Society for Microbiology.
Antimicrobial Resistance. World Health Organization.
Bacteriological Analytical Manual. U.S. Food and Drug Administration.
Antimicrobial Susceptibility Testing. Clinical and Laboratory Standards Institute.
Culture Collections. American Type Culture Collection.
NCBI Pathogen Detection. National Center for Biotechnology Information.
Genomic Epidemiology, Evolution, and Transmission Dynamics of Porcine Deltacoronavirus.. Molecular biology and evolution, 2020.
Expansion and transmission dynamics of high risk carbapenem-resistant Klebsiella pneumoniae subclones in China: An epidemiological, spatial, genomic analysis.. Drug resistance updates : reviews and commentaries in antimicrobial and anticancer chemotherapy, 2024.
Untangling lineage introductions, persistence and transmission drivers of HP-PRRSV sublineage 8.7.. Nature communications, 2024.
Tracing the Evolutionary History and Global Expansion of Candida auris Using Population Genomic Analyses.. mBio, 2020.

This article is educational and does not replace validated laboratory procedures, clinical diagnosis, institutional biosafety review, or public-health guidance.