Ofir Raz

Acute myeloid leukemia (AML) is associated with unfavorable patient outcomes primarily related to disease relapse. Since specific types of leukemic hematopoietic stem and progenitor cells (HSPCs) are suggested to contribute to AML propagation, this study aimed to identify and explore relapse-initiating HSPC subpopulations present at diagnosis, using single-cell analysis (SCA). We developed unique high-resolution techniques capable of tracking single-HSPC-derived subclones during AML evolution. Each subclone was evaluated for chemo-resistance, in vivo leukemogenic potential, mutational profile, and the cell of origin. In BM samples of 15 AML patients, GMP-like and MLP-like HSPC subpopulations were identified as prevalent at relapse, exhibiting chemo-resistance to commonly used chemotherapy agents cytosine arabinoside (Ara-C) and daunorubicin. Reconstruction of phylogenetic lineage trees combined with genetic analysis of single HSPCs and single-HSPC-derived subclones demonstrated two distinct clusters, originating from MLP-like or GMP-like subpopulations, observed both at diagnosis and relapse. These subpopulations induced leukemia development ex vivo and in vivo. Genetic SCA showed that these relapse-related subpopulations harbored mutated EZH2 and TP53, detected already at diagnosis. This study, using combined molecular, functional, and genomic analyses at the level of single cells, identified patient-specific chemo-resistant HSPC subpopulations at the time of diagnosis, promoting AML relapse.

Whole-genome amplification is a crucial first step in nearly all single-cell genomic analyses, with the following steps focused on its products. Bias and variance caused by the whole-genome amplification process add numerous challenges to the world of single-cell genomics. Short tandem repeats are sensitive genomic markers used widely in population genetics, forensics, and retrospective lineage tracing. A previous evaluation of common whole-genome amplification targeting ~1000 non-autosomal short tandem repeat loci is extended here to ~12,000 loci across the entire genome via duplex molecular inversion probes. Other than its improved scale and reduced noise, this system detects an abundance of heterogeneous short tandem repeat loci, allowing the allelic balance to be reported. We show here that while the best overall yield is obtained using RepliG-SC, the maximum uniformity between alleles and reproducibility across cells are maximized by Ampli1, rendering it the best candidate for the comparative heterozygous analysis of single-cell genomes.

Short tandem repeats (STRs) are polymorphic genomic loci valuable for various applications such as research, diagnostics and forensics. However, their polymorphic nature also introduces noise during in vitro amplification, making them difficult to analyze. Although it is possible to overcome stutter noise by using amplification-free library preparation, such protocols are presently incompatible with single cell analysis and with targeted-enrichment protocols. To address this challenge, we have designed a method for direct measurement of in vitro noise. Using a synthetic STR sequencing library, we have calibrated a Markov model for the prediction of stutter patterns at any amplification cycle. By employing this model, we have managed to genotype accurately cases of severe amplification bias, and biallelic STR signals, and validated our model for several high-fidelity PCR enzymes. Finally, we compared this model in the context of a naïve STR genotyping strategy against the state-of-the-art on a benchmark of single cells, demonstrating superior accuracy.

Cell lineage analysis aims to uncover the developmental history of an organism back to its cell of origin ¹ . Recently, novel in vivo methods and technologies utilizing genome editing enabled important insights into the cell lineages of animals ^2–8 . In contrast, human cell lineage remains restricted to retrospective approaches, which still lack in resolution and cost-efficient solutions. Here we demonstrate a scalable platform for human cell lineage tracing based on Short Tandem Repeats (STRs) targeted by duplex Molecular Inversion Probes (MIPs). With this platform we accurately reproduced a known lineage of DU145 cell lines cells ⁹ and reconstructed lineages of healthy and metastatic single cells from a melanoma patient. The reconstructed trees matched the anatomical and SNV references while adding further refinements. Our platform allowed to faithfully recapitulate lineages of developmental tissue formation in cells from healthy donors. In summary, our lineage discovery platform can profile informative STR somatic mutations efficiently and we provide a solid, high-resolution lineage reconstruction even in challenging low-mutation-rate healthy single cells.

Advances in biochemical technologies have led to a boost in the field of single cell genomics. Observation of the genome at a single cell resolution is currently achieved by pre-amplification using whole genome amplification (WGA) techniques that differ by their biochemical aspects and as a result by biased amplification of the original molecule. Several comparisons between commercially available single cell dedicated WGA kits (scWGA) were performed, however, these comparisons are costly, were only performed on selected scWGA kit and more notably, are limited by the number of analyzed cells, making them limited for reproducibility analysis. We benchmarked an economical assay to compare all commercially available scWGA kits that is based on targeted sequencing of thousands of genomic regions, including highly mutable genomic regions (microsatellites), from a large cohort of human single cells (125 cells in total). Using this approach, we could analyze the genome coverage, the reproducibility of genome coverage and the error rate of each kit. Our experimental design provides an affordable and reliable comparative assay that simulates a real single cell experiment. Results demonstrate the needfor a dedicated kit selection depending on the desired single cell assay.

Short tandem repeats (STRs) are polymorphic genomic loci valuable for various applications such as research, diagnostics and forensics. However, their polymorphic nature also introduces noise during in vitro amplification, making them difficult to analyze. Although it is possible to overcome stutter noise by using amplification-free library preparation, such protocols are presently incompatible with single cell analysis and with targeted-enrichment protocols. To address this challenge, we have designed a method for direct measurement of in vitro noise. Using a synthetic STR sequencing library, we have calibrated a Markov model for the prediction of stutter patterns at any amplification cycle. By employing this model, we have managed to genotype accurately cases of severe amplification bias, and biallelic STR signals, and validated our model for several high-fidelity PCR enzymes. Finally, we compared this model in the context of a naïve STR genotyping strategy against the state-of-the-art on a benchmark of single cells, demonstrating superior accuracy.

The field of synthetic biology is fueled by steady advances in our ability to produce designer genetic material on demand. This relatively new technological capability stems from advancements in DNA construction biochemistry as well as supporting computational technologies such as tools for specifying large DNA libraries, as well as planning and optimizing their actual physical construction. In particular, the design, planning, and construction of user specified, combinatorial DNA libraries are of increasing interest. Here we present some of the computational tools we have built over the past decade to support the multidisciplinary task of constructing DNA molecules and their libraries. These technologies encompass computational methods for [1] planning and optimizing the construction of DNA molecules and libraries, [2] the utilization of existing natural or synthetic fragments, [3] identification of shared fragments, [4] planning primers and overlaps, [5] minimizing the number of assembly steps required, and (6) correcting erroneous constructs. Other computational technologies that are important in the overall process of DNA construction, such as [1] computational tools for efficient specification and intuitive visualization of large DNA libraries (which aid in debugging library design pre-construction) and [2] automated liquid handling robotic programming [Linshiz et al., Mol Syst Biol 4:191, 2008; Shabi et al., Syst Synth Biol 4:227–236, 2010], which aid in the construction process itself, have been omitted due to length limitations.