RT @elezzx2020: Excited to present our work on developing jaxQTL, a new single-cell eQTL mapping tool that improves power and robustness in….

RT @CharlestonCWKC: It's already a busy 2025, and we're still in January! Batting first from the lab, led by @JiTang1024, we present the as….

@CharlestonCWKC Many thanks to @bldinh and Jalen Langie for providing valuable datasets/tools for this study. Also many thanks to everyone in the lab for bringing so much joy into this journey.(13/13).

Many thanks to my postdoc advisor @CharlestonCWKC for all the wonderful guidance. You always give me so brilliant advice on questions relevant to and not relevant to research. Genuinely speaking super lucky to be in your lab and you are my role model! (12/13).

We also used other simulated and empirical datasets to evaluate as-eGRM, the conclusions are generally the same. Please see the pre-print for more info. (11/13).

We observed that as-eGRM is still able to reveal the structure and outperforms the alternatives, reflecting as-eGRM is more robust to missing data.(10/13)

We next performed the evaluation using all the individuals from the Chicago center across the entire ancestry proportion spectrum. The task is harder as a lower proportion of the target ancestry segments means more missing data due to masking non-target ancestry segments. (9/13).

We further evaluated as-eGRM using the SOL data. We first followed previous studies to use only the sampless with the proportions of the segments from target ancestry > 0.5. With these samples, as-eGRM successfully replicates the structures observed by the previous studies.(8/13)

When the proportions of the segments from the non-target ancestry are between 0.3~0.5, as-eGRM outperforms the alternatives. Increasing the proportions to 0.4~0.6(the task is harder as greater admixture reduces the portion of the informative genome), asegrm is more robust.(7/13)

When the timing of the split to establish the structure is 300 generations ago, as-eGRM outperforms the alternatives. With decreasing the timing to 100 (the task is harder as a more recent structure means less differentiation among the sub-pops), asegrm is more robust.(6/13)

Firstly, using the data simulated by a grid-like stepping stone model, we evaluated how the proportions of the segments from the non-target ancestry as well as the timing of the split to establish the structure impact as-eGRM and the alternatives.(5/13).

We extensively evaluated as-eGRM and compared it to the alternatives with multiple simulated and empirical datasets.(4/13).

We developed a method,as-eGRM, leveraging estimated genealogical trees and local ancestry calls across the genome to estimate the genetic relationship matrix within ancestry components to reveal ancestry-specific structures in admixed populations via dimensional reduction. (3/13).

Existing methods for elucidating these structures generally rely on frequency-based estimates of pairwise genetic relatedness and disregard linkage information between markers, potentially limiting their resolution.(2/13).

Genetic admixtures are pervasive in human populations. Elucidating ancestry-specific structures in admixed populations is crucial for comprehending population history and mitigating confounding effects in GWAS.(1/13)

Excited to share my first preprint from @CharlestonCWKC lab. We developed a genealogy-based method for revealing ancestry-specific structures in admixed populations. Interested in this topic? Please read the thread🧵, we'd love to hear your thoughts and feedback!.

RT @biorxiv_genetic: A genealogy-based approach for revealing ancestry-specific structures in admixed populations .

RT @uscpphs: 🌟Charleston Chiang, PhD, has been promoted to Associate Professor of Population and Public Health Sciences. Chiang who current….