Xue Li
Research Scientist, MPI Core B
Texas Biomedical Research Institute
I am a computational biologist specialized in statistical, bioinformatic and population genetic data processing and analysis. My major research focus is using large scale whole-genome sequencing data to understand the parasite population genomics (such as spread of resistance alleles under intense selection, and patterns of parasite transmission) during malaria elimination efforts in Eastern Myanmar. For malaria parasite genetic studies, I led the adaptation of bulk segregant analysis (BSA) approaches in malaria parasite Plasmodium falciparum, including development of efficient experimental methodology and bioinformatic pipelines for quantitative trait loci (QTL) identification. We have been using this BSA method to localize genomic regions that determine parasite fitness, drug resistance and nutritional genetics. I developed analysis approaches for accurate genotyping parasite genomes from Illumina short-read data as well as from Nanopore long-read data, and applied those approaches to maximize the utility of malaria parasite genetic crosses.
Research Interests
Application of bulk segregant analysis in parasite biology
Classical linkage mapping allows detailed examination of the genetic architecture of phenotypic traits, including interactions between causative genes, and is the method of choice for many parasite traits. However, this approach with malaria parasite involves cloning, sequencing, and phenotyping multiple independent progeny clones, which is laborious and expensive. An alternative approach, bulk segregant analysis (BSA), uses pools of recombinant progeny, and quantitative trait loci (QTL) are identified by exposing progeny pools to selection and examining skews in allele frequency across the genome. BSA has several advantages over the classical approaches for selectable traits: it is faster and more cost-efficient; statistical power is increased because pools containing 1000s of recombinants can be examined and it narrows the width of QTL peaks, because more recombination events are sampled 1.
I led the adaptation of BSA approaches in malaria parasite P. falciparum, including development of efficient experimental methodology and bioinformatic analysis pipelines to identify QTL regions. Highlights include using the BSA method to localize genomic regions that determine parasite fitness across the whole life cycle 2, and identification of genes underlying drug resistance 3, 4 or associated with parasite nutrient acquisition/metablolism 5. Notably, by combining BSA approach, longitudinal field data from Africa, and CRISPR/cas9 gene editing, we demonstrated that pfaat1 (an amino acid transporter gene in P. falciparum) plays a key role in chloroquine (CQ) resistance evolution 4. PfAAT1 is involved in both parasite fitness and CQ resistance, and had strikingly different trajectories of resistance evolution in Asia and Africa. Our work shows the power of BSA approach, which also led to my collaboration with Prof. Manoj Duraisingh to localize genes involved with merozoite invasion and host species preference. I am also working closely with Prof. Boris Striepen and Dr. Sebastian Shaw to pioneer use of BSA in Cryptosporidium parasites.
Using nanopore long-reads to maximize the utility of malaria genetic crosses
P. falciparum has an AT-rich, haploid genome, with highly variable subtelomeric regions that contain polymorphic gene families involved in immune evasion and pathogenesis. I have developed bioinformatic pipelines for accurate genotyping of P. falciparum genome from short read sequences 6, and applied the analysis to characterize parents and progeny from genetic crosses 2, 4, 7, 8. We have parents and 1497 progeny (686 independent recombinants) Illumina sequenced from three malaria parasite genetic crosses. However, analysis using short-read sequencing technologies is limited to the core genome regions, and cannot resolve rearrangements and complex structural variants. To expand the utility of Plasmodium genetic crosses, we used Nanopore long-read technology to sequence parental parasites and five progeny from each of the three genetic crosses. We performed de novo assembly to generate highly contiguous, telomere-to-telomere genome sequences for both parents and progeny.
By mapping Illumina short reads to parental reference sequences, we were able to expand the percentage of genome sequenced, improve resolution of QTL peaks, and resolve gene content in subtelomeric regions. By comparing parental and progeny assembly, we uncovered segregating and de novo structural variants (SVs) and rearrangements from each cross, which include SVs in genes such as Rh2a/b, pfmdr1, GTP cyclohydrolase 1, MSP11 and in genes from the Apicomplexan AP2 (ApiAP2) family. We also identified rapid evolution of plasmepsin II/III (pfpm2/3) copy number variants (CNVs) in cross Mal31xKH004 8, which is associated with piperaquine (PPQ, a widely used partner drug for dihydroartemisinin [DHA] in malaria treatment) resistance. These results demonstrate the value of long sequencing reads and high-quality genome assembly for malaria genomic analysis.
Structural variation in malaria parasites
The malaria parasite P. falciparum has a great capacity for evolutionary adaptation to evade host immunity and develop drug resistance. However, ~37% of the P. falciparum genome is either highly repetitive or highly variable, and thus difficult to analyze using short-read sequencing technologies 6, 9, leading to critical gaps in our understanding of their evolution and association with phenotypes. I have developed a long-read analysis approach to identify complex variations from parents and progeny of genetic crosses. I plan to use these approaches to improve the understanding of genomic diversity evolution in P. falciparum genome from the following two aspects.
Resolving parasite genome dynamics during meiotic recombination. Recombination is a key biological process that generates and maintains genetic diversity in natural parasite populations and thus contributes to parasite evolution. I will use both Nanopore long-read and Illumina short-read sequencing data from the genetic crosses described above to investigate rate and placement of chromosomal breakpoints from meiosis, and sequence signatures at recombination hotspots. I will compare these recombination signatures with the distribution and characteristics of SVs identified above, to investigate the mutational mechanisms giving rise of these variations. I am also interested in studying the phenotypic consequences of SVs generated through recombination. For example, we used progeny clones carrying different copies of pfpm2/3 to directly test the impact of pm2/3 CNV on drug resistance related phenotypes 8. We also identified CNVs and structural variations in progeny at GTP cyclohydrolase I (pfgch1), the first and rate limiting enzyme in the folate biosynthesis pathway. Higher pfgch1 copy numbers compensate for less fit mutations in pfdhfr and pfdhps, however the question on how parasites adjust flux through the folate pathway remained. I plan to use IBD progeny with different form of pfgch1 to explore this parasite’s ability to maintain a balance of flux. This study will be important to development of alternative agents targeting this crucial synthesis pathway.
Polymorphisms of clonally variant genes (CVGs) during asexual multiplication. CVGs include gene families such as var, rif, stevor, eba or ap2, that are associated with pathogenesis, antigenic variation, erythrocyte invasion or sexual conversion in P. falciparum. Our discovery of structural variation in ap2 family genes from genetic cross data is of particular interest, give the important role of this gene family in many aspects of Plasmodium biology. How CVG diversity is generated in asexual parasites is not currently completely understood, while structure rearrangements were identified around var genes during mitosis 10-12. We have two parasite isolates growing in human erythrocytes in vitro for over 200 days, with cultures cryopreserved and DNA materials collected every week. I will long-read sequence clones from this collection and analyze CVG dynamics such as mutation rates and patterns. I will also characterize short-read mapping signatures of the SVs identified, and thus investigate their distribution in population datasets.
Understanding malaria parasite evolution in low transmission settings
Regions of low malaria transmission intensity predominate in Southeast (SE) Asia and South America and are now becoming increasingly common in Africa 13. A central challenge for malaria control is to develop efficient approaches to eliminate malaria from such regions. My research goal is to understand parasite transmission, population genomics and resistance evolution in those regions, and use these data to inform future control efforts.
Understanding the rapid expansion of kelch13-R561H lineage in Kayin State, Myanmar. In our previous study, we characterized the rapid expansion of a kelch13-R561H lineage in Kayin State, Myanmar. We have two major hypotheses: 1) the parasites are more fit or more resistant to the anti-malaria drugs used in Kayin State; 2) the parasites are more likely to escape immune reactions by generating more asymptomatic infections. To test hypothesis 1, we will measure resistance level (e.g. IC50s) of different kelch13-R561H clones, and perform head-to-head competition experiments to quantify parasite fitness. We will also generate experimental crosses between parasites that differ in drug resistance and fitness traits. We will map the location of the genes and their mutations that control these phenotypes using BSA selections. We will further use CRISPR/cas9 gene editing to validate whether any of the mutations identified play a role in parasite resistance or fitness. To test hypothesis 2, we will combine long-read genomic sequencing and RNAseq to identify immune escape polymorphisms by comparing samples collected from symptomatic and asymptomatic infections. Parasite escapes immune responses built from pre-existing host antibodies by regularly switching the expression of its ~60 var genes. The long-read sequencing approaches described above allow us to analysis this kind of variations.
Genomic evolution in clonal transmitted parasites. We identified high level of clonal parasite transmission in Kayin State, Myanmar, with 9 lineages (43-229 clones sequenced per lineage) sampled over 36 months. Long-read sequencing of these lineages will allow us to investigate how complex variations were generated and selected in nature at the host-pathogen interface. I will focus on one of the lineages with the longest duration time (over 50 months). Mutations that have arisen and spread to fixation within the clonal linage are particular interesting, as they may contribute to parasite resistance evolution or fitness restoration. This dataset will allow me to characterize mutations involved with lineage differentiation, which can also be used for targeted surveillance on Thailand-Myanmar border.