Table 2.

Studies combining phylogenetic and social network analysis to characterize transmission networks (N = 20)

Authors	Title	Pub. yeara	Study design	Locationb	Time period	Data source/project	Population characteristics	Social/sexual network data	Molecular cluster analysis	Data integration	Resultsc
Campbell et al.	Phylodynamic Analysis Complements Partner Services by Identifying Acute and Unreported HIV Transmission	2020	Cohort	North Carolina, New York City, San Francisco	2011–2013	Screening Targeted Populations to Interrupt Ongoing Chains of HIV Transmission with Enhanced Partner Notification (STOP) Project	1326 cohort members with newly diagnosed acute HIV infection. 94.8% male, 65.0% non-white, median age of 30 years, 78% MSM	Enhanced partner notification was offered to participants who screened positive at one of 12 participating HIV testing sites	Partial HIV pol sequences from persons with available blood specimens (70 participants with AHI and 547 total participants) were analyzed with Microbe-Trace using genetic distance threshold of d< 1.5%	Genetic networks and sexual contact networks were constructed independently and then integrated into a multi-partite network	12.6% of HIV positive participants (n = 1326) had an Acute HIV Infection. Among participants with available pol sequences, 465 high-risk contacts were reported, 65.7% of which were genetically supported as a transmission pair. Five links between individuals with AHI were reported, but none were genetically supported. Phylodynamic analysis identified an additional 102 unreported putative transmission links. Participants who were in a cluster were more likely to be younger. Black men who met partners online were most likely to be in the largest clusters
Deng et al.	Molecular Epidemiological Tracing of HIV-1 Outbreaks in Hainan Island of Southern China	2009	Cohort	Hainan Island Province, China	1991–2006	Chinese national HIV/AIDS surveillance system and sentinel surveillance program	499,752 individuals from high-risk groups (IVDUs, female sex workers & their clients, STD clinic attendees, truck drivers, pregnant women, MSM, TB patients, former plasma donors, and children of HIV-infected mothers)	Participants completed standardized questionnaire administered by the Hainan Center for Disease Control & Prevention that included demographics, risk factors, mode of transmission, and sexual contacts. A second round of contact tracing was conducted on individuals found to be in the molecular clusters that had not previously completed a questionnaire	Blood samples were collected from a representative subset of all participants based on HIV prevalence in each county (n = 88). Phylogenetic relationships between viral pol sequences were estimated using neighbor-joining trees and maximum likelihood models	Phylogenetic tree was used to confirm or reject possible transmission events suggested by contact tracing data	523 of the 499,752 participants were diagnosed with HIV infection. The 83 available patient samples segregated into one large and three small clusters. Phylogenetic analysis confirmed four purported heterosexual transmission links suggested by contact tracing and rejected two purported links
Dennis et al.	Social Network Based Recruitment Successfully Reveals HIV-1 Transmission Networks Among High Risk Individuals in El Salvador	2013	Cohort	San Salvador, San Miguel, Sonsonate; El Salvador	2008	Encuesta Centroamericano de Vigilancia de VIH y Comportamiento en Poblaciones Vulnerables (ECVC)	699 MSM and 757 Female Sex Workers (FSW)	HIV-positive MSM and female sex workers were recruited via multiples waves of respondent driven sampling, which was used to construct a social network. Recruits were interviewed about their demographics, sex practices, and history of STIs	A phylogenetic tree was constructed from pol sequences from 119 HIV-positive individuals recruited via respondent driven sampling. Mean pairwise genetic distance <0.015 was used to identify transmission clusters	The phylogenetic network was compared to the network built from respondent-driven sampling	Of the available sequences, 43% of MSM and 10% of FSW had viral sequences that grouped into 14 transmission clusters. Clusters were grouped by risk factors (12 MSM clusters) and by geography. Only two transmission clusters had participants directly linked through respondent recruitment. All clusters greater than two individuals spanned more than one RDS chain. Cluster membership was associated with recent HIV infection, sex with a stable male partner, and sex with three or more partners in the past year
Fujimoto et al.	Lack of Support for Socially Connected HIV-1 Transmission Among Young Adult Black Men Who Have Sex with Men	2017	Cohort	Houston, TX	2015	Integrated Molecular & Affiliation Network analysis of HIV transmission (iMAN)	10 study participants. All Black MSM aged 17–29, average age of 25	Young black MSM were recruited via respondent driven sampling. A peer-referral network based on social and sexual connections was constructed from an adjacency matrix which was then maximally symmetrized. Two venue-based networks among the 10 individuals were also constructed	Clinical samples from subset of 10 iMAN participants were analyzed. Models of sequence evolution were chosen using Akaike’s Information Criteria. Maximum likelihood and Bayesian phylogenetic analyses were performed using AIC-chosen models. GenBank was used to identify 20 control HIV-1 sequences	Sequences from individuals found to be connected in RDS-derived network were analyzed for genetic similarity to assess for evidence of transmission pairs	No evidence was found to suggest transmission of viruses between any individuals in the socially-connected RDS chain. Support for relationships among viruses sampled from different individuals were low, thus did not conform to the structure expected for a recent transmission cluster. One viral sample was not placed in the clade with the other case samples, but instead was most closely related to a GenBank control sequence
Grande et al.	Transmission Patterns in a Low HIV-Morbidity State—Wisconsin, 2014–2017	2019	Cohort	WI	2014–2017	Wisconsin Department of Public Health	1401 individuals with molecular sequences reported to Wisconsin DPH during 2014–2017. Characteristics only provided for the 433 individuals in molecular clusters (88.4% male, 56.1% Black, 80.8% MSM, 3.5% MSM + IDU)	Data from partner services interviews were obtained from Wisconsin’s Partner Services Web	HIV-1 sequence data from persons with HIV infection in Wisconsin were analyzed using HIV-TRACE. Sequences considered related if genetic distances were <0.015. Transmission clusters defined as group of greater than two linked sequences	Named linkages between sexual contacts were compared to molecular linkages to assess whether named partners had highly genetically similar sequences	30.9% of HIV sequences reported to Wisconsin DPH during 2014–2017 had molecular linkage to at least one other person. People were most commonly molecularly linked to persons of their own racial group, sexual orientation, and age range, except for Hispanics/Latinos and persons aged 13–19 years old. 33.8% of named partner linkages identified in public health interviews also had a molecular linkage
Kostaki et al.	Molecular Analysis of Human Immunodeficiency Virus Type 1 (HIV-1)–Infected Individuals in a Network-Based Intervention (Transmission Reduction Intervention Project): Phylogenetics Identify HIV-1-Infected Individuals With Social Links	2018	Cohort	Athens, Greece	2013–2015	Transmission Reduction Intervention Project (TRIP)	356 participants 18+ years old comprised of HIV+ “seeds” (both recent and chronic infections as two discrete recruitment arms), their social network recruits, and HIV negative controls. 42.1% HIV+, 90.2% IDU	Social network data was obtained through study recruitment links, network questionnaires, and field/venue observations. 1st and 2nd degree social contacts of HIV+ seeds were recruited to the study. Contacts were defined as people who had sex with or injected with/in the presence of a seed (1st degree) or a first-degree contact of the seed (2nd degree). If a newlydiagnosed individual was identified among network recruits, two steps of contact tracing was then performed with this new seed	Samples were obtained for 118 out of 150 HIV+ study participants. Subtypes were identified using the online COMET tool and confirmed using 247 globally sampled reference sequences. Phylogenetic trees were constructed using maximum likelihood, neighbor joining, and Bayseian methods. Transmission clusters (2 or more sequences) were considered “highly supported” if they met confidence criteria for all three methods	The proportion of individuals with social network ties within their corresponding transmission clusters was assessed. These findings were compared to the expected proportion and distribution found in simulated random social mixing	13 highly supported transmission clusters with 2–5 individuals each were generated by phylogenetic analysis. 43.8% of individuals had social network ties within their own clusters. Hypothesis testing indicated that this was significantly higher than the expected proportion with random mixing and significantly different from the distribution of expected proportions
Lee et al.	An Exploratory Study on the Social and Genotypic Clustering of HIV Infection in Men Having Sex with Men	2009	Cohort	Hong Kong	2007–2008	Hong Kong Integrated Treatment Centre (HIV specialist clinic)	73 MSM receiving care at the Hong Kong Integrated Treatment Center clinic who were diagnosed with HIV in the preceding year, 18+, and of Chinese ethnicity. Mean age 33.5 years	Participants completed a questionnaire that collected sociodemographic data, estimated year of HIV infection, the type and number of sexual partners, and physical venues/internet sites used to seek sexual partners. Likert scales assessed frequency of venue usage and sexual behaviors. Venue-based clusters were constructed using Likert score strength	HIV-pol genes from 49 patients were sequenced and aligned using Clustal X. A phylogenetic tree was constructed using the maximum likelihood method with 1000 bootstrap replications. Clusters were defined as 3+ closely related sequences	Venue-based clusters constructed via social network analysis were compared to phylogenetic clusters, and genetic relatedness within venue clusters was assessed	Out of three identified phylogenetic clusters, two featured participants who mainly sourced partners from the internet (9/10 and 5/6 participants). Those in the third phylogenetic cluster mainly sourced partners from saunas and not from the internet. The sauna-based sourcing cluster also had viral sequences that were less genetically related to one another than the internet-based cluster
Lepej et al.	Phylogenetic Analysis of HIV Sequences Obtained in a Respondent-Driven Sampling Study of Men Who Have Sex with Men	2009	Cohort	Zagreb, Croatia	2006	Croatian Ministry of Health	360 MSM in Zagreb recruited via respondent-driven sampling. 100% MSM	Zagreb-based MSM were recruited via multiples waves of respondent driven sampling, which was used to construct a social network. Participants were asked about their sexual behaviors and tested for various STIs	Sequences were obtained from HIV+ study participants. Reference sequences were obtained from 20 unrelated Croatians through routine clinical monitoring. Phylogenetic analysis was performed on pol sequences using the neighbor-joining method with the Kimura two-parameter distance model. Bootstrap analysis was used to assess branching pattern reliability, with clusters above 98% considered significant	HIV Molecular clusters were compared with the respondent-driven sampling recruitment chain to determine the relationship between both data sources and help characterize HIV transmission within the cohort	Out of 360 total participants, 18 (5%) were HIV+. Sequencing was successfully performed on 12 (3.3%) samples. Five of the sequences clustered together. HIV+ phylogenetically related cases were recruited in later waves: 1 in the 4th wave, 2 in the 6th, and 1 in the 8th. This phenomenon suggests the role of a sexual network in HIV transmission, though it appears unlikely that participants recruited their sexual partners
Lin et al.	Behavioral and molecular tracing of risky sexual contacts in a sample of Chinese HIV-infected men who have sex with men	2013	Cohort	Taizhou Prefecture, China	2008–2010	Chinese National Information System for AIDS Prevention and Control	100 MSM index cases newly diagnosed with HIV between 2008 and 2010 in Taizhou Prefecture. Mean age 30.3 482 sexual contacts identified via contact tracing. 89% MSM	Index cases filled out egocentric contact tracing surveys and identified sexual partners were offered HIV testing. Sexual partners testing HIV+ were also asked to fill out egocentric contact tracing surveys. This process was repeated until no new HIV cases were identified	Sequences of gag and env from 72 HIV+ participants were compared with control sequences and aligned using Clustal X. Phylogenetic and molecular evolutionary analyses were conducted. Evolutionary distances were calculated and phylogenetic dendrograms were constructed using the neighbor-joining method	Identification of potential HIV transmission pairs was based on both sexual and molecular connections between the pair. A transmission cluster was defined as a group with at least 1 potential HIV transmission pair	51 of the 100 HIV+ index cases were newly identified through contact tracing. Index cases together reported 1534 sexual contacts and provided contact information for 482 (31%), of which 115 (24%) ultimately received HIV testing. 7 out of 49 independent sexual networks were deemed HIV transmission clusters and were supported by phylogenetic analysis of HIV sequences. 16 networks included HIV transmission pairs linked by contact tracing that were not supported genetically
Monroe-Wise et al.	Peer-Mediated HIV Assisted Partner Services to Identify and Link to Care HIV-Positive and HCV-Positive People who Inject Drugs: a Cohort Study Protocol	2021	Cohort	Nairobi, Kilifi, Mambosa Counties; Kenya	2018—Ongoing	Kenya Medical Research Institute, Kenya Ministry of Health	1000 HIV + IDU index cases enrolled across 8 study sites and their sex/drug use partners identified via partner services	Index cases complete questionnaires identifying all of their sex and drug use partners over the last three years. Clinicians and peer educators then engage community partners for study enrollment, partner services/demographics questionnaire completion, and rapid HIV/HCV testing	HIV and HCV sequencing will be attempted for all study participants at the KwaZulunatal Research Innovation and sequencing Platform. Study HIV sequences will be combined with publicly available molecular data prior to performing molecular cluster analysis	Sexual/drug-use partner and molecular cluster data will be utilized to characterize viral transmission among high-risk populations and identify traits associated with increased infectivity	Trial currently at preresults stage
Morgan et al.	Sexual, Social, and Genetic Network Overlap: A Socio-Molecular Approach Toward Public Health Intervention of HIV	2018	Cohort	Chicago, II	2013–2016	uConnect Cohort	266 HIV+ Black MSM aged 16–29 living in South Side Chicago. 69% identified as gay, 57% reported condom-less sex in the last 12 months	Participants were recruited via respondent driven sampling. All participants completed surveys that collected information on sociodemographics, substance use, close confidants, and recent sexual partners. Consenting participants (n = 12) also provided lists of their Facebook friends. Researchers used this data to construct recruitment, confidant, sexual, and Facebook networks	HIV pol sequences (n = 86) were obtained from all persons whose viral load was ≥ 2000 copies/mL through a combination of dried blood spot collection and surveillance data from the Chicago Department of Public Health. Phylogenetic tree analyses were performed by using the neighbor-joining method. Specimens were linked if the genetic distance between pol sequences was less than or equal to 0.015 substitutions/ site. Clusters were defined as 2 or more specimens linked by 1 + ties	Molecular, sexual, and confidant networks were consecutively integrated to postulate likely routes of HIV transmission. All analyses were restricted to named partners who were also study participants	35 HIV sequences (41%) formed clusters with 55 total molecular ties. There were no significant differences in demographics, sexual activity, drug use, or confidant, sexual, and Facebook network ties between clustered and unclustered sequences. None of the molecular ties were identified by 1st, 2nd, or 3rd degree confidant or sexual networks, or 1st degree recruitment networks. There was a consistent 45–50% overlap in confidant, sexual, Facebook, and recruitment networks
Pasquale et al.	Leveraging Phylogenetics to Understand HIV Transmission and Partner Notification Networks	2018	Cohort	Wake County, NC	T2012–2013	NC Department of Health and Human Services (NC-DHHS), North Carolina Screening and Tracing for Active HIV-1 Transmission (NC-STAT) program	280 index cases, defined as newly diagnosed HIV cases in state of NC during 2012–2013. 83% male, 65% Black, 40% under 30 years old. 383 sexual partners or high-risk social contacts	Data on demographics, HIV testing history, HIV lab results, and sexual and social contacts were abstracted from DIS interviews performed by NC-DHHS or Wake County DHHS for all persons in NC newly diagnosed with HIV. A Partner Notification network was constructed using named partner and high-risk contact data	HIV-1 pol sequences sampled between 1997 and 2014 were obtained from the largest reference lab in NC (n = 15,246). Sequences that did not match an index case or contact were considered background sequences. Sequences were aligned using MUSCLE and phylogenetic trees were constructed using maximum-likelihood phylogenies. Putative transmission clusters were defined as clades with high branch support, maximum pairwise distance less than 3.5% between all sequences, and inclusion of at least one index or partner case	Index and partner cases were matched with HIV sequences using date of birth, gender, and lab test dates. Separate partner notification and transmission cluster networks were created and differences between the two networks were compared	80% of index cases were interviewed, reporting a total of 854 sex partners and 34 social contacts. Of these partnerships, 383 were unique non-index partners and 335 were located during investigation. 34% of partners were found to be HIV+, while 27% had an unknown HIV status. 73% of individuals in the partner notification network who had sequences were in one of the 116 transmission clusters identified. 59% of transmission clusters contained links between sexual network components that were not apparent from the partner notification network
Pilon et al.	Transmission Patterns of HIV and Hepatitis C Virus among Networks of People Who Inject Drugs	2011	Cohort	Ottowa, Canada	2007	Ontario HIV Treatment Network	407 IDUs enrolled from September to December 2007 consisting of 7 ‘seeds’ and their recruits. 100% IDUs, 80% Male, 10.1% HIV+	7 Initial IDU ‘seeds’ were selected based on participation in previous studies. Seeds were asked to recruit 3 IDU peers via a chain-referral method. New recruits were also instructed to recruit 3 additional peers, with this process continuing until a target of 400 participants was reached	HIV pol and HCV sequences were amplified via RTPCR. Sequences were aligned via Clustalx software, and phylogenetic trees were constructed using Saitou and Nei’s neighbor-joining method with the Kimura 2-Parameter model. Bootstrapping was used to identify significant clusters using a threshold of 80% reliability when resampling	Chain-referral social networks and phylogenetic data were compared to decipher if and how social relationships influenced transmission dynamics	Out of 407 total recruits, 41 (10.1%) were HIV+, of which 40 (9.9%) were HIV/HCV coinfected. 29 (71%) HIV samples were ultimately sequenced, 18 (62%) of which fell into one of 7 distinct clusters ranging from 2 to 4 participants. Two of the unique clusters contained participants separated by 2 or less recruitment cycles, while the remaining 5 clusters contained more
Pines et al.	Concurrency and HIV Transmission Network Characteristics among MSM with Recent HIV Infection	2016	Cohort	San Diego, CA	1996–2015	San Diego Primary Infection Resource Consortium (SD-PIRC) & Center for AIDS Research Network of Integrated Clinical Systems (CNICS)	986 HIV+ individuals recruited through: (1) The UCSD HIV screening centers between 1996 and 2005, aged 16+, recently HIV diagnosed, and ART-naive (SD-PIRC cohort, n = 800), (2) The UCSD HIV clinic between July 2007 and August 2013 (CNICS cohort, n = 186). Characteristics provided for subgroup of 285 recently infected, cisgender MSM in SD-PIRC cohort used in partner concurrency analysis (White 60%, 24% Hispanic, median age 33 years)	Individuals in the SD-PIRC cohort completed interviews on socio-demographics, drug use, sexual behavior, and partner-specific data on up to three recent sexual partners. Partner concurrency was assessed based on reported overlapping timelines of sexual partnerships within 3 months prior to interview completion	HIV-1 pol sequences were collected from both cohorts (n = 986). HIV TRACE was used to infer partial transmission networks. Putative transmission links were inferred when the genetic distance between two sequences was less than 1.5%	Data on the dates of HIV diagnosis, sample collection, and estimated dates of HIV infection were examined for putative transmission links to assess direction of transmission and whether transmission occurred during a recent infection	In the subgroup of recently HIV infected MSM (n = 285), 54% phylogenetically clustered with others in the network by one or more putative transmission links. 54% reported having concurrent partners. There was a positive association between concurrency and transmission network clustering and network degree
Resik et al.	Limitations to Contact Tracing and Phylogenetic Analysis in Establishing HIV Type 1 Transmission Networks in Cuba	2007	Cohort	Cuba	Not identified	Not identified	127 HIV-1 seropositive individuals, both male and female, identified as members of two discrete contact tracing networks. HIV seroconversion of network members occurred predominantly from 1982 to 1993	Contact tracing was used to identify one network of 38 HIV+ individuals and another of 89, with an emphasis on partners at the time of HIV diagnosis	Blood samples were obtained from consenting study participants. RTPCR was used to amplify the gag, env1, and env2 genomic portions of viral samples. Phylogenetic trees were constructed using a maximum likelihood heuristics approach. Subtyping was performed using reference sequences from GenBank	Phylogenetic analysis was performed among samples in each contact tracing network and was used to verify or negate possible transmission links	In network 1 (N = 38), phylogenetic analysis of gag supported 4 transmission events, rejected 6, and did not reject 2. Analysis of env supported 2 transmission events, rejected 3, and did not reject 9. Combined env/gag analysis supported 3 transmission events and did not reject 2. In network 2 (N = 89), combined env/gag analysis supported 5 transmission events and did not reject 3
Smith et al.	A Public Health Model for the Molecular Surveillance of HIV Transmission in San Diego, California	2009	Cohort	San Diego County, CA	2006–2007	San Diego County Resistance Testing program (SDC cohort), First Choice Program (FCP)	268 HIV+ individuals recruited through referrals to drug resistance testing (SDC cohort). 89% male, 62% Caucasian, mean age 35 years, 80% MSM, 7% MSM + IDU. 369 recruited through the Acute Infection and Early Disease Research Program (FCP cohort) from 1996 to 2007. 96% male, 79% Caucasian, 92% MSM; 36 sexual partners recruited through contact tracing of the FCP cohort. 92% male, 62% Caucasian, 91% MSM	Contact tracing was performed with members of the FCP cohort to elicit contact information of sexual partners. Epidemiologically linked partners were defined as sexual partners with available contact information who the index cases believed to be the likely source of their HIV infections	Phylogenetic analysis was performed on HIV-1 pol sequences obtained from all SDC and FCP cohort members and listed sexual partners using BioEdit and MUSCLE. Transmission clusters were defined when pol sequences from any two people were >99% genetically similar	Phylogenetic analysis was used to evaluate the genetic ties between index cases and their epidemiologically linked partners identified through contact tracing	In 673 total participants, pol sequences were on average 5.1% genetically different. 25% of participant sequences clustered when combining the SDC and FCP cohorts, with the largest cluster consisting of 12 participants. 24 of the 36 (67%) epidemiologically linked partners were supported by phylogenetic analysis. Epi-linked partner sequences clustered with 60% of the clustering FCP and SDC sequences and were grouped into 15 clusters
Schneider et al.	Abstract: Do Partner Services Initiated from Molecular Clusters Yield New or Viremic HIV Cases?	2020	Cohort	Chicago, IL	2012–2016	Chicago Department of Public Health (CDPH)	1015 newly diagnosed HIV+ index cases reported to the CDPH from 2012 to 2016 with available HIV sequences. Characteristics for the 336 index cases with phylogenetically clustered sequences: Average age 28, 47% Black, 29% Latinx, 89% MSM	Newly diagnosed HIV+ index cases were offered partner services by CDPH at time of diagnosis. Partner notification data was extracted from CDPH records	Molecular clusters were constructed using HIV-TRACE at a pairwise genetic distance threshold of 0.5%	Integration of partner notification data and phylogenetic analysis was used to assess the difference in yields between clustered and nonclustered index cases and identifying sex/drug-use partners with new HIV diagnoses or viremia at time of diagnosis	Out of 1015 index cases, 336 (33%) had HIV sequences that clustered. 96 index cases named a total of 539 sex/drug-use partners. Out of all listed partners, 162 (37%) were linked to clustered index cases. 20% of the listed partners were either new HIV diagnoses or viremic at the time of diagnosis. There was no significant different in the yield of new or viremic HIV partners linked to clustered vs. nonclustered index cases
Ssemwanga et al.	HIV Type 1 Subtype Distribution, Multiple Infections, Sexual Networks, and Partnership Histories in Female Sex Workers in Kampala, Uganda	2012	Cohort	Kampala, Uganda	2008–2009	Good Health for Women Project (GHWP)	324 HIV+ women who were a part of a larger cohort of 1027 high-risk sex and local entertainment facility workers in Kampala enrolled between March 2008 and April 2009 who also provided blood samples	At 3 months intervals, participants completed sociodemographic and behavioral questionnaires. Information on women’s’ life stories and sexual partnerships was also elicited during a study featuring a subset of the cohort (n = 200) who were found to have phylogenetically similar viruses and be at high risk of contracting STIs	HIV-1 gp-41 and pol sequences were obtained from 210 participants. Reference sequences from the Los Alamos Sequence database were used, and Clustal X was used to align the sequences. Phylogenetic trees were constructed using neighbor joining method. Clusters with bootstrap values > 90% were considered phylogenetically similar	Life stories were used to determine when/where sex work was performed. The location of sex work was compared for women with phylogenetically similar viruses	In the 210 participants, HIV subtype A (~60%) was the most prevalent subtype. Sexual networks of 6 pairs and one triplet of participants with similar viral sequences were identified. Four dyads of women with phylogenetically similar viruses shared similar sex work venues. Five cases of multiple infections (9%) were also discovered
Tordoff et al.	Combining Traditional and Molecular Epidemiology Methods to Quantify Local HIV Transmission among Foreign- Born Residents	2021	Cohort	King County, WA	2010–2018	HIV/STD Program, Public Health – Seattle & King County, National HIV Surveillance Program	2409 HIV+ residents of King County, Washington diagnosed between January 1st, 2010, and December 31st, 2018. 33% foreign-born	Partner services was offered to all newly diagnosed HIV+ residents of King County as defined by the National HIV Surveillance System. All interviews were conducted by Public Health—Seattle & King County	HIV-1 pol sequences were obtained for 1448 study participants. Distance-based clustering analyses were used to identify clusters of 2+ individuals using Tamura-Nei pairwise genetic distance with a threshold of 0.02 substitutions per site. Sensitivity analyses were also conducted using thresholds of 0.025, 0.015, and 0.01 substitutions per site. A phylogenetic tree was constructed using FastTree approximate maximum likelihood method	Molecular cluster and partner services data were combined to infer the location of HIV acquisition for foreign-born individuals	Among individuals with available sequences (n = 1448), 1104 (76%) clustered into 295 genetically similar clusters. Combined partner services and molecular data were used to infer the HIV acquisition location for 611 (77%) of all 798 HIV+ foreign born residents, with 254 (32%) presumably acquiring HIV outside King County, 394 (49%) outside of the US, 205 (26%) in King County, and 13 (2%) elsewhere in the US. Partner services data alone inferred HIV acquisition locations for 100 (13%) participants, and molecular cluster data inferred acquisition locations for 258 (32%) participants
Wertheim et al.	Social and Genetic Networks of HIV-1 Transmission in New York City	2017	Cohort	New York, NY	2006–2012	New York City Department of Health and Mental Hygiene	756 HIV+ index cases who were reported to the NYC Department of Public Health. 52% Black, 41% Hispanic, 45% MSM, 7% IDUs; 586 HIV+ sex and IDU partners of the index cases	HIV+ index cases were interviewed by the NY Field Services Unit for the names of their sex and IDU partners over the previous 12 months. Named partners were contacted and referred to testing and care. Index cases and named partners were classified by their primary transmission risk factor (i.e. IDU)	HIV-1 sequences were obtained for index cases and HIV+ named partners (n = 1342) through the NYC Department of Health. HIV-TRACE was used to construct the genetic transmission network. Sequences were aligned to the HXB2 reference sequence and pairwise Tamura-Nei 93 genetic distances between sequences were calculated. Pairs of sequences falling below a range of distance thresholds were then linked, with connected components forming transmission clusters	Molecular and sexual/IDU network data were used to create two separate and one merged HIV transmission network. These networks were used to determine a range of genetic distance thresholds suitable for identifying potential transmission partners as well as to determine factors associated with genetic linkage in possible transmission partners	Based on 756 index cases and 586 HIV+ identified partners, plausible thresholds for genetic distance were calculated to be between 0.01 and 0.02 substitutions/site. 43–65% of genetic linkage clusters were supported by contact tracing data while 37–49% of contact tracing networks were supported by genomic analysis. Genomic data supported HIV transmission along 310 edges between 651 partner names in the contact tracing network, while contact tracing supported 388 out of 736 edges in the genomic network. Contact tracing did not identify possible transmission in over half of HIV cases across distance thresholds. 449 out of 756 (59%) index cases were genetically linked to at least one named partner. Bidirectionally named partners had increased odds for being genetically linked than unidirectionally named partners

^aYear of publication

^bCity and/or state abbreviation

^cMSM Men who have sex with men, HIV human immunodeficiency virus, IDU injection drug user, CDC centers for disease control and prevention, BLAST basic local alignment search tool, BRAI bio-rad avidity incidence, IVDU intravenous drug user, STD sexually transmitted disease, TB tuberculosis, ECVC Encuesta Centroamericano de Vigilancia de VIH y Comportamiento en Poblaciones Vulnerables, FSW female sex worker, STI sexually transmitted infection, Hx history; NC-DHHS North Carolina Department of Health and Human Services, NS-STAT North Carolina Screening and Tracing for Active HIV-1 Transmission, AHI acute HIV infection, RHI recent HIV infection, DIS disease intervention specialists; iMAN integrated molecular & affiliation network analysis of HIV transmission, AIC Akaike’s Information Criterion, RDS respondent-driven sampling, DPH department of public health, TRACE secure HIV transmission cluster engine, PLWHA people living with HIV/AIDS, PCR polymerase chain reaction; PWID people who inject drugs; COMET context-based modeling for expeditious typing, GRT genotypic resistance testing, CDPH Chicago Department of Public Health, MUSCLE multiple sequence comparison by log-expectation, HCV hepatitis C virus, SD-PIRC San Diego Primary Infection Resource Consortium, CNICS center for AIDS research network of integrated clinical systems, RTPCR reverse transcription polymerase chain reaction, AIEDRP acute infection and early disease research program, FCP first choice program, ART antiretroviral therapy, GHWP good health for women project, UCSD University of California San Diego, SDC San Diego County, NYCDOH New York City Department of Health