Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Oct 4.
Published before final editing as: ACS Chem Biol. 2026 Sep 11:10.1021/acschembio.6c00605. doi: 10.1021/acschembio.6c00605

Ultramutagenesis through combined Msh2 depletion and proofreading-deficient DNA polymerase ε expression enhances small molecule target identification in forward genetic screens

Alexander-Hoi Nguyen a,b, Vicky Li a, Wei-Min Chen c,d, Jacob Kimberg a,b, Anthony J Davis c,d, David G McFadden a,b,c,d,*, Juan Manuel Povedano a,b,d,*
PMCID: PMC13633561  NIHMSID: NIHMS2209994  PMID: 42723219

Abstract

DNA mismatch repair (MMR) deficiency has been widely utilized in forward genetic screens to identify drug resistance mutations and elucidate the mechanisms of action of cytotoxic small molecules. However, MMR deficiency generates a characteristic mutational signature enriched for C>T transitions that precludes saturation of mutagenesis. We hypothesized that expression of proofreading deficient DNA polymerase epsilon, as has been observed in patients with biallelic MMR deficiency, would enhance mutagenesis and enable identification of new resistance mutations and cytotoxin mechanisms. Here, we combine auxin-inducible degradation of Msh2 with doxycycline-inducible expression of the proofreading-deficient PolεP286R mutant in murine small cell lung cancer (SCLC) cells. Simultaneous MMR deficiency and mutant Polε expression markedly increased the emergence of bortezomib-resistant clones relative to either perturbation alone and targeted sequencing of Psmb5 in bortezomib-resistant clones identified recurrent mutations previously reported in chemical mutagenesis-based resistance screens. We next applied this platform to investigate resistance to lurbinectedin, a clinically relevant therapy for SCLC. Whole-exome sequencing revealed recurrent loss-of-function alterations in nucleotide excision repair (NER) genes, most notably Ercc5 and Ercc4, implicating NER deficiency as a major mechanism of lurbinectedin resistance and supporting a central role for NER in mediating lurbinectedin-induced cytotoxicity. Collectively, these findings establish inducible ultramutagenesis as a powerful and versatile platform for the unbiased discovery of drug resistance mechanisms and therapeutic vulnerabilities.

Graphical Abstarct

graphic file with name nihms-2209994-f0001.webp

INTRODUCTION

The identification of the molecular mechanism of action of cytotoxic small molecules is a critical step in drug development. Knowledge of the direct target of a small molecule provides insight into target engagement, on-target effects, and potential mechanisms of resistance. Genetic approaches to elucidate the mechanism of action commonly rely on the identification of mutations that confer resistance by disrupting the interaction between a small molecule and its protein target1, 2. Such strategies require the generation of sufficient genetic diversity to enable the emergence and selection for resistant clones. Historically, elevated mutation rates have been achieved through chemical mutagenesis or by utilizing cell lines with defects in mismatch repair (MMR), which exhibit increased spontaneous mutagenesis 2–5. These approaches have been successfully employed to identify resistance-conferring alleles and uncover the molecular targets of cytotoxic compounds. However, the spectrum of mutations generated by MMR deficiency is biased, potentially limiting the identification of resistance.

One approach to generating MMR-deficient cancer cell lines is through genetic inactivation of MSH2 or the use of inducible protein degradation systems targeting essential MMR components5, 6. We and others have previously demonstrated that MMR deficiency facilitates the acquisition of resistance-conferring mutations to a variety of cytotoxic small molecules 5–10. Despite these successes, the published literature likely overestimates the overall efficiency of this strategy. In our experience, only a subset of small molecules yields resistant clones in MMR-deficient cells, highlighting important limitations of MMR-based mutagenesis approaches. Several factors may contribute to the low success rate. First, some small molecules may not exert their effects through a discrete protein target, thereby hindering the emergence of target-based resistant mutations. Second, resistance-conferring mutations must disrupt compound activity while preserving the essential physiological function of the target protein, a constraint that may substantially limit the number of viable resistance alleles in essential proteins. Third, the mutational spectrum generated by MMR deficiency is enriched for C>T transitions and may not generate the full spectrum of potential resistance mutations.

Biallelic mismatch repair deficiency (bMMRD) is a rare hereditary cancer predisposition syndrome caused by germline loss-of-function mutations in core MMR genes. Defective MMR promotes the rapid accumulation of somatic mutations and frequently leads to the acquisition of secondary mutations in the proofreading (exonuclease) domains of DNA polymerases ε (Polε) and δ (Polδ). The combined impairment of DNA repair and replication fidelity drives exceptionally high mutation rates, termed “ultramutation,” resulting in tumors with mutation frequencies that exceed 250 mutations per megabase11. This natural occurring synergy between MMR deficiency and defective polymerase proofreading provides a compelling model for engineering extreme mutator phenotypes.

In this study, we engineered inducible systems for Msh2 degradation and expression of a proofreading-deficient Polε mutant into cancer cell lines to enhance the spectrum of mutations capable of generating small molecule resistance. Using this platform, we screened for resistance to two cytotoxic small molecules and successfully identified recurrent mutations in genes encoding proteins required for their cytotoxic activity. These findings establish inducible ultramutagenesis as an effective strategy for forward genetic screens aimed at elucidating drug mechanisms of action and resistance pathways.

RESULTS

We hypothesized that combination of MMRd and expression of exonuclease deficient Polε would increase and broaden the spectrum of mutations in cancer cell lines. To generate an inducible MMR-deficient model, we used CRISPR-Cas9 to engineer an auxin-inducible degron (AID) tag into the endogenous Msh2 locus of murine SCLC (518T2) cells (Figure 1A)12. Following blasticidin selection, single-cell clones were isolated, expanded, and screened for Msh2 expression by western blot analysis (Figure 1B). We identified a clone that lacked detectable wild-type Msh2 protein and instead expressed a higher-molecular-weight species consistent with successful integration of the AID tag at the endogenous Msh2 locus (Figure 1B).

Figure 1. Engineering an auxine-inducible degron into the Msh2 locus in murine SCLC cells.

Figure 1.

Open in a new tab

(A) Schematic depicting the knock-in of an AID tag followed by a P2A and blastocidin resistant cassette (BSD). (B) Western-blot showing the expression of Msh2 (left) and the Msh2-AID (right) from murine SCLC cells and engineeered Msh2-AID cells, respectively. α-Tubulin was used as loading control. Independently assessed one time. (C) Western-blot of lentiviral transduced murine SCLC cells showing the expression of the plant Cullin adapter protein, TIR1, using an antibody recognizing the V5 tag. Gapdh was used as loading control. Independently assessed one time. (D) Western-blot showing the degradation of Msh2-AID in a dose-dependent manner at 24h. Actin was used as loading control. Assay was independently conducted three times, one independent well at each concentration is presented.

To enable inducible degradation of AID-tagged Msh2, we introduced the auxin-responsive E3 ubiquitin ligase adaptor protein TIR1F74A via lentiviral transduction and confirmed stable expression by immunoblot analysis using a V5 antibody (Figure 1C and Figure S1). The TIR1F74A variant exhibits enhanced affinity for the synthetic auxin analog 5-phenyl-indole-3-acetic acid (5-Ph-IAA), allowing efficient target protein degradation at substantially lower concentrations13, 14. Treatment of Msh2-AID;TIR1F74A cells with 150–200 nM 5-Ph-IAA induced rapid and efficient degradation of Msh2, with complete loss of detectable protein observed within 24 hours (Figure 1D). Collectively, these results established a robust and tunable system for the inducible degradation of endogenous Msh2 in murine SCLC cells, enabling temporal control of MMR deficiency.

Next, we used a PiggyBac transposon-based vector to introduce a Polε transgene harboring the P286R exonuclease-domain mutation, which impairs proofreading activity and increases replication-associated mutagenesis11, 15, 16. The transgene was placed under the control of a doxycycline-inducible Tet-On system containing a TRE3G promoter, enabling temporal regulation of PolεP286R expression (Figure 2A).

Figure 2. Tetracycline inducible promoter for mutant Polε overexpression in murine SCLC cells.

Figure 2.

Open in a new tab

(A) Schematic depicting the tetracycline-inducible expression of catalytically dead polymerase ε, leading to mutagenesis. (B) Western-blot showing the expression of Polε in a dose-dependent manner at 48 hours in TREG-PolεP286R SCLC cells. Gapdh used as a loading control. Assay was independently conducted three times, one independent well at each concentration is presented. Modified from BioRender. Povedano selfa, J. M. (2026) https://BioRender.com/3mjhxtc

Treatment of 518T2 cells with non-toxic concentrations of doxycycline resulted in robust induction of PolεP286R expression within 48 hours, as confirmed by immunoblot analysis (Figure 2B and Figure S2). These engineered cells enabled independent or combined manipulation of two complementary mutagenic pathways: (1) induction of MMR deficiency through Msh2 degradation following 5-Ph-IAA treatment, (2) expression of proofreading-deficient PolεP286R following doxycycline treatment, (3) simultaneous induction of both perturbations through combined treatment with 5-Ph-IAA and doxycycline (Figure 3A). This experimental system provided a versatile platform to evaluate the individual and combined contributions of MMR deficiency and defective polymerase proofreading to mutagenesis and the emergence of drug-resistant clones.

Figure 3. Emergence of resistant clones increases under ultramutator condition.

Figure 3.

Open in a new tab

(A) A schematic for forward genetics screening using 518T2Ultra. (B) Resazurin viability assay for 518T2Ultra cells treated with indicated concentrations of bortezomib. Red dotted line indicates Median Z-Score = 5. Analyzed one time. (C) Number of clones detected across each condition, obtained from three independent selections at three different concentrations (7.71 nM, 8.20 nM, 8.72 nM). *p<0.05, **p<0.01, ****p<0.0001 Pearson’s chi-squared test. (D) Ratio of bortezomib EC50 of selected clones to parental 518T2Ultra, dotted red line indicates ΔEC50 = 1.5. (E) Ratio of CD437 EC50 of selected clones to parental 518T2Ultra, dotted red line indicates ΔEC50 = 1.5. (D-E) ● represents clones with mutations found in Psmb5. ● represents clones where Psmb5 mutations were not found.

To determine whether combined MMR deficiency and PolεP286R expression (hereafter referred to as MMRd-PolεP286R) enhanced the emergence of drug-resistant clones relative to either perturbation alone, we performed proof-of-principle resistance screens using bortezomib, a proteasome inhibitor with a well-characterized target (Psmb5) and established resistance-conferring mutations17, 18. To establish selection conditions, we first determined the minimum bortezomib concentrations required to achieve complete cell killing after one week of drug exposure. We then performed resistance selections at 7.24, 7.71, 8.2, and 8.72 nM bortezomib in untreated control cells, 5-Ph-IAA treated MMR-deficient cells, doxycycline treated PolεP286R-expressing cells, and 5-Ph-IAA/doxycycline-treated MMRd-PolεP286R cells (Figure S3A). Across all bortezomib concentrations tested, the non-mutagenic condition (DMSO) produced the fewest resistant clones. In contrast, MMRd-PolεP286R cells consistently yielded a significantly greater number of bortezomib-resistant clones than either MMR-deficient or PolεP286R expressing cells alone (Figure 3B-C).

Resistant clones from each condition were isolated, expanded, and subjected to dose-response analyses to confirm bortezomib resistance. To evaluate whether resistance resulted from a compound-specific mechanism rather than a generalized multidrug-resistant phenotype, we also assessed sensitivity to CD437, a DNA polymerase α inhibitor (Figure 3C-D and Figure S3B-I).

Clones isolated from the DMSO control condition exhibited bortezomib sensitivities comparable to those of parental Msh2-AID;TIR1F74A murine SCLC cells (EC50 fold-change <1.5), indicating that these colonies did not represent bona fide resistant clones. In contrast, all clones derived from MMR-deficient or PolεP286R-expressing cells were resistant to bortezomib while remaining sensitive to CD437 (EC50 ≤ 1.5-fold higher than parental cells). Similarly, 17 of 18 clones isolated from the MMRd-PolεP286R condition exhibited robust bortezomib resistance and retained sensitivity to CD437 (Figure S3C-J).

To identify the genetic basis of resistance, we performed Sanger sequencing of Psmb5. Twenty-eight of 32 resistant clones harbored previously described resistance-associated mutations in Psmb5 (Figure S3L), including 7/9 clones derived from the PolεP286R condition and 16/18 derived from the MMRd-PolεP286R condition4, 5, 19. Notably, we identified the Psmb5C133W mutation, a known resistance allele previously reported in chemical mutagenesis screens that has not been observed in screens using MMR-deficient cells 4. To directly test whether this allele was sufficient to confer resistance in our system, we introduced the C133W substitution into parental 518T2 cells using CRISPR-Cas9-mediated homology-directed repair. Cells were nucleofected with a Cas9-sgRNA ribonucleoprotein complex together with an HDR donor oligonucleotide encoding the Psmb5C133W mutation. Two days after nucleofection, cells were treated with 15 nM bortezomib for 10 days, after which surviving cells were allowed to recover for an additional week. Resistant outgrowth was observed by crystal violet staining only in the nucleofected condition, whereas control non-nucleofected cells failed to recover (Figure S3M). These results supported the conclusion that targeted CRISPR-Cas9-mediated introduction of the Psmb5C133W allele conferred bortezomib resistance in parental 518T2 cells.

Together, these findings demonstrate that combining Msh2 depletion with expression of proofreading-deficient Polε significantly enhances the emergence of bortezomib-resistant clones. The high frequency of recurrent Psmb5 mutations supports the notion that the increased frequency of resistant clones in the MMRd-PolεP286R condition was driven by enhanced mutagenesis of the ultramutator system.

We next applied the ultramutator platform to investigate mechanisms of resistance to lurbinectedin, a DNA-damaging agent approved for the treatment of extensive-stage SCLC as maintenance therapy following first-line treatment20. Lurbinectedin binds covalently to guanine residues within GC-rich regions of the DNA minor groove, generating DNA adducts that are processed by the transcription-coupled nucleotide excision repair (TC-NER) pathway. During this repair process, TC-NER-mediated excision of lurbinectedin-DNA adducts results in the formation of DNA double-strand breaks, leading to activation of the DNA damage response and ultimately apoptosis.21–24 To identify genetic determinants of lurbinectedin resistance, we performed forward genetic screens using five lethal concentrations of lurbinectedin in ultramutator (300 nM 5-Ph-IAA and 1 μg/mL doxycycline) and mutagenesis-off control (DMSO) Msh2-AID;TIR1F74A murine SCLC cells. Following selection, resistant clones were isolated and expanded for further characterization Across selections performed at lurbinectedin concentrations ranging from 161 to 259 pM, we recovered nine resistant clones (Figure 4A), seven of which were determined to be independent based on unique DNA barcodes introduced prior to selection (Lurb-A through Lurb-G). Dose-response analyses confirmed that all seven independent clones were resistant to lurbinectedin, exhibiting greater than two-fold increase in EC50 relative to the DMSO-treated parental ultramutator line. Importantly, these clones retained sensitivity to bortezomib, indicating that resistance was specific to lurbinectedin rather than the result of a generalized drug-resistant phenotype (Figure 4B-D).

Figure 4. Forward genetics screening using 518T2Ultra reveals the mechanism of resistance for lurbinectedin.

Figure 4.

Open in a new tab

(A) Resazurin viability assay for 518T2Ultra cells treated with indicated concentrations of lurbinectedin. Median Z-Score for wells with oversaturated fluorescence were calculated with RFU=100,000. Independently assessed one time. (B) CellTiter-Glo Viability assay for lurbinectedin-resistant clones treated with lurbinectedin for 3 days. (C) CellTiter-Glo Viability assay for lurbinectedin-resistant clones treated with bortezomib for 3 days. (B-C) Each viability assay was independently conducted one time with two technical duplicates at each concentration, error bars represent std error of the mean between technical duplicates. (D) EC50 for lurbinectedin and bortezomib for each lurbinectedin-resistant clone. (E) Ercc5 is mutated in 4 out of 7 clones. (F) WES analysis results for lurbinectedin-resistant clones reveals mutations in ERCC4 and ERCC5. (G) Western blot of lurbinectedin-resistant clones for Ercc5. α-Tubulin was used as a loading control. (H) Clonogenic survivability assay measures radiation sensitivity of 518T2Ultra, Lurb-D, and Lurb-G. Assay independently repeated three times, shown is the relative survival between technical duplicates of a single independent replicate, error bars represent std error of the mean between technical duplicates.

Whole-exome sequencing of lurbinectedin-resistant clones (Lurb-A through Lurb-F) identified recurrent mutations in four genes (Ercc5, Spata31h1, Ttn, and Xirp) (Figure 4E). Among these candidates, Ercc5 was particularly notable, as multiple clones harbored frameshift mutations predicted to result in loss of function. This finding is consistent with previous studies demonstrating that ERCC5 deficiency conferred resistance to lurbinectedin (Figure 4F).23 To determine whether these genetic alterations affected Ercc5 protein expression, we performed immunoblot analysis. Clones Lurb-B through Lurb-F exhibited complete loss of detectable Ercc5 protein, whereas Lurb-A displayed substantially reduced expression (Figure 4G). Interestingly, no coding mutations in Ercc5 were identified in Lurb-A or Lurb-F despite diminished protein expression, suggesting that non-coding alterations may have driven the loss of Ercc5 in these clones. In addition, Lurb-B, Lurb-C, and Lurb-E each harbored a frameshift mutation in only one Ercc5 allele, while the second allele appeared wild type by exome sequencing. Despite this, Ercc5 protein was undetectable in all three clones (Figure 4G). In contrast, Lurb-D contained two independent Ercc5 mutations, consistent with the observed complete loss of protein expression. Together, these findings suggested that additional genetic alterations, potentially located in non-coding regulatory regions or structural variants not captured by exome sequencing, contributed to reduced or absent Ercc5 expression observed in lurbinectedin-resistant clones, including Lurb-A, Lurb-B, Lurb-C, Lurb-E, and Lurb-F.

Reduced or absent Ercc5 expression accounted for lurbinectedin resistance in all resistant clones except Lurb-G, which retained ERCC5 expression comparable to that of parental cells. Notably, whole-exome sequencing revealed a homozygous frameshift mutation in Ercc4 in the Lurb-G clone (Figure 4F). ERCC4 encodes XPF, a critical component of the nucleotide excision repair (NER) pathway. Loss of several NER factors, including XPG (ERCC5), XPF (ERCC4), XPA, and XPD (ERCC2), was previously shown to confer resistance to ecteinascidin-743, the parent compound of lurbinectedin.23 These observations suggested that Lurb-G also harbored an NER defect despite retention of ERCC5 expression.

Because suitable antibodies for murine Ercc4 were unavailable, we sought to functionally assess DNA repair capacity in Lurb-G cells. Defects in NER are known to confer hypersensitivity to DNA-damaging agents, including ultraviolet (UV) radiation, due to impaired repair of bulky DNA lesions.25 Therefore, we compared the sensitivity of parental 518T2 Msh2-AID;TIR1F74A murine SCLC cells, the Ercc5-deficient clone Lurb-D, and the Ercc4-mutant clone Lurb-G following exposure to increasing doses of UV radiation. After irradiation, cells were allowed to recover and viability was assessed. Both Lurb-D and Lurb-G exhibited markedly increased sensitivity to UV radiation relative to parental 518T2 Msh2-AID;TIR1F74A cells, consistent with impaired NER function (Figure 4G-H and Figure S4). Together, these findings demonstrated that the lurbinectedin-resistant clones identified in our screen harbored defects in NER, arising through loss-of-function alterations in either Ercc5 or Ercc4. These results further supported the central role of NER in mediating the cytotoxic effects of lurbinectedin.

To evaluate whether the combined MMRd/Polε-mutant background alters the mutational spectrum beyond an MMRd-only state, we performed SBS96 mutational-signature analysis using WES data from lurbinectedin-resistant clones. We analyzed somatic single-base substitutions with variant allele fraction (VAF) > 0.20 in the resistant clone and VAF < 0.01 in the parental cell line, classified each mutation according to its trinucleotide context, and assigned COSMIC SBS signatures using SigProfilerAssignment. The lurbinectedin-resistant ultramutator clones exhibited heterogeneous fitted signature activities across clones, including MMR-deficiency-associated components such as SBS6, SBS15, and SBS21, together with additional components including SBS1, SBS5, SBS17a, SBS54, SBS59, and SBS106 (Figure S4D). Because matched lurbinectedin-resistant clones from MMRd-only and Polε-mutant-only conditions were not available, we used previously published WES data from Msh2-KO murine SCLC clones that were generated by selection with a microtubule destabilizing compound8. This comparison showed that the Msh2-KO cells exhibited a more restricted spectrum dominated by T>C and T>G substitutions, with lower C>T contribution (Figure S4E). In contrast, the ultramutator-derived lurbinectedin-resistant clones showed clone-to-clone variation in SBS96 profiles, including increased C>T contribution in several clones and detectable representation of additional substitution classes (C>A, C>G, and T>A) (Figure S4E)8. These results were consistent with the conclusion that combining MMR deficiency with Polε-mutant expression produces heterogeneous mutational outputs less restricted to a canonical MMRd-associated spectrum.

DISCUSSION

Here, we engineered an inducible ultramutator system in cancer cell lines by combining MMR deficiency through Msh2 degradation with expression of a proofreading-deficient Polε mutant. Simultaneous induction of Polε mutant and Msh2 degradation resulted in a marked increase in the emergence of bortezomib-resistant clones relative to either MMR deficiency or mutant Polε expression alone, consistent with an elevated mutation rate following combined loss of MMR and expression of an exonuclease-deficient DNA polymerase (Polε). These findings mirror observations from biallelic MMR-deficient tumors, which frequently acquire secondary mutations in POLE or POLD1 and subsequently develop exceptionally high mutational burdens. 11 Similarly, mouse models harboring proofreading-deficient Polε alleles develop a broad spectrum of tumor types with elevated mutational burden, further supporting the central role of defective polymerase proofreading in driving accelerated mutagenesis and tumor evolution.15, 16 Together, these observations suggest that combining MMR deficiency with impaired polymerase proofreading recapitulates a naturally occurring mechanism of hypermutation and provides a powerful experimental platform for enhancing mutagenesis and genetic diversity in forward genetic screens.

The inducible ultramutagenesis observed in our system enabled the identification of resistance-associated mutations to clinically relevant compounds including lurbinectedin. Similar forward genetic approaches have previously been used to define mechanisms of drug resistance2, 3, 5, 7, 10. However, the elevated mutation rate generated by our model may allow more comprehensive sampling of mutational space, including a broader range of codons within drug targets as well as additional genes that contribute to resistance mechanisms. Supporting this idea, we identified the Psmb5C133W mutation, which had not previously been reported in MMR-deficient models but was identified in chemically induced mutagenesis screens 4, 26. In addition, the lurbinectedin-resistant alleles identified in this study were exclusively loss-of-function frameshift mutations, in contrast to prior reports that predominantly identified missense mutations. Moreover, the ultramutator cell line enabled the emergence of loss-of-function resistant mutations requiring biallelic inactivation, as observed in lurbinectedin clone D. This ultramutator strategy may therefore facilitate the identification of resistance mechanisms for compounds whose targets are less tractable by conventional approaches, including drugs that do not act through a single protein target. Together, these findings suggest that inducible ultramutagenesis may uncover resistance mechanisms that were inaccessible using conventional mutagenesis approaches.

Although our results do not identify a direct protein target of lurbinectedin, our results are consistent with previous studies demonstrating that an intact NER pathway is required for lurbinectedin cytotoxicity23. Notably, no coding mutations in Ercc5 were detected in the Lurb-F clone despite complete loss of Ercc5 protein expression. This observation raises the possibility that resistance may arise through noncoding regulatory mutations that suppress Ercc5 expression or through loss-of-function alterations in factors required for Ercc5 transcriptional regulation, mRNA processing, or protein stability. Transcriptional stalling of RNA polymerase II (RNA Pol II) has been proposed as a key mediator of lurbinectedin-induced toxicity.27 However, we did not identify mutations in RNA Pol II subunits in any of the resistant clones. While the absence of such mutations does not exclude a role for transcriptional stalling in lurbinectedin activity, it argues against a model in which resistance commonly arises through direct alteration of RNA Pol II. Moreover, given the essential role of RNA Pol II in cellular viability, mutations that disrupt drug binding while preserving normal polymerase function may be intrinsically rare. Instead, our findings support a model in which disruption of NER represents the predominant mechanism of lurbinectedin resistance. The recurrent loss-of-function alterations identified in Ercc5 and Ercc4, together with the established functional coupling between NER and transcription, suggest that impairment of NER-dependent processing of lurbinectedin-induced DNA lesions is sufficient to attenuate the cytotoxic effects of the drug.

In conclusion, we engineered an inducible ultramutator system in murine cancer cells by combining targeted degradation of Msh2 with expression of a proofreading-deficient Polε mutant. This approach yielded substantially more clones than MMR deficiency alone, thereby increasing the recovery of resistance-conferring mutations and facilitating the identification of drug resistance mechanisms. Our findings establish combined MMRd and proofreading deficient polymerase induction as a powerful platform for forward genetic screens aimed at elucidating drug mechanisms of action and resistance pathways.

METHODS

Cell Culture

Murine SCLC cell line, 518T2, was cultured at 37°C with 5% CO2 in DMEM (Sigma #D6429) supplemented with 5% FBS (Corning 35-010-CV) and 2mM L-glutamine (Corning 25-005-CL).

Western Blotting

Western-blotting was performed using standard methods using pre-cast gels (Invitrogen NW04125BOX). Bio-Rad Nitrocellulose Membranes (Bio-Rad, #1620112) were used for protein transference and then blocked using Bio-Rad Blotting-Grade Blocker in TBS (Bio-Rad, #1706404) for 1 hour at RT. Primary antibodies used were anti-V5 tag (D3H8Q) (Cell Signaling Technologies #13202), anti-GAPDH (GeneTex GTX627408), anti-PolE (GeneTex GTX132100), and anti-Msh2 [D24B5] XP rabbit mAb (#2017, Cell Signaling Technology). Primary antibodies were incubated for 1 hr at RT diluted 1:1000 in Bio-Rad Blotting-Grade Blocker: TBS-Tween (0.1%). Membrane was washed with TBS-Tween (0.1%) three times for 5 minutes each wash. Secondary antibodies used were IRDye 800CW Donkey anti-Mouse (#925–32212, LI-COR), IRDye 800CW Donkey anti-Rabbit (#926–32213, LI-COR), IRDye 680RD Donkey anti-Rabbit (#926–68073, LI-COR), and IRDye 680RD Donkey anti-Mouse (#926–68072, LI-COR). Visualization was performed with Odyssey CLx Imaging System (LI-COR).

Compounds

Bortezomib was purchased from Selleck Chemicals (#S1013). CD437 was purchased from Sigma-Aldrich (#C5865). Lurbinectedin was purchased from MedChemExpress (#HY-16293). Compounds were resuspended in DMSO (Fisher Bioreagents #BP231–100) at a concentration of 10mM in aliquots of 100 μL and kept at −20C.

Dose-Response Curves

Cells were plated in 96-well plates, 10,000 cells per well in 200μL of media. After overnight incubation, compounds were dispensed using a D300e Digital Dispenser (TECAN). Cell viability was assessed after 72 hours using CellTiter-Glo luminescent cell viability assay (Promega, #G7571). CellTiter-Glo reagent was diluted by adding PBS-Triton-X (1%) (1:1 ratio). Luminescence was assessed using a cytation 5 plate reader (BioTek).

Selection of Resistant Clones

Selection of bortezomib resistant clones was performed by plating cells under DMSO, 300nM 5-Ph-IAA, 1μg/mL Doxycycline, or a combination of 300nM 5-Ph-IAA and 1μg/mL Doxycycline and grown for 3wks. Cells were then plated in 96 well plates, 10,000 cells per well in 200uL of media. After overnight incubation to allow cells to adhere, bortezomib was added using a D300e Digital Dispenser (TECAN), and cells were cultured in the presence of compound for 14 days. To identify surviving clones, cells were cultured in 0.04 mg/ml resazurin (BioRad) in DMEM without phenol red, supplemented with 5% FBS, 1% penicillin/streptomycin, and 1% L-glutamine. Fluorescence was measured on a Cytation 5 plate reader (BioTek) with a 560/20 excitation filter and a 590/20 emission filter to identify wells with surviving cells. Clones identified with resazurin were verified by visual inspection, and the media of verified clones were changed to DMEM (Sigma D6429) supplemented with 5% FBS and 2mM L-glutamine. The median Z-score was determined as Median Z Score = ((Individual Well Fluorescence Intensity – Median Fluorescence Intensity per mutagenesis condition)/(Median Absolute Deviation of Individual Well Flourescence)). Surviving clones were expanded in standard growth medium.

We classified wells with a median fluorescence Z-score >5 as clone-positive, containing a compound resistant clone, and wells with a Z-score ≤5 as clone-negative.

Identification of Psmb5 mutations by Sanger sequencing

Genomic DNA was isolated from bortezomib and lurbinectedin resistant clones using Zymo Research Quick-DNA Miniprep (Zymo Research #D3024). The exon 2 of Psmb5 was amplified using the gDNA from bortezomib resistant clones as template and the following primers: 5’ -AAA AAG ATG CAA AGC CAG GTT – 3’ and 5’ - GGC TCA CAG GAC ACA ACT CA – 3’. The resulting amplicon was subjected to Sanger sequencing using 5’ - GTA TGG GGA GTA TTT GTG GTC TTA CGG – 3’.

Psmb5 C133W Point Mutation

We performed homology-directed repair using the Alt-R CRISPR-Cas9 system and an Ultramer ssODN donor from Integrated DNA Technologies (IDT) in 518T2 mSCLC cells. To prepare the gRNA complex, we combined 5 μL of Alt-R CRISPR-Cas9 crRNA (100 μM; sequence: 5′- UCU GGG AGC GGU UGU UGG CUG UUU UAG AGC UAU GCU-3′) and 5 μL Alt-R CRISPR-Cas9 tracrRNA (100 μM). The mixture was heated to 95°C for 5 minutes and then allowed to slowly cool to room temperature. The ribonucleoprotein complex was formed by combining 9 μL of gRNA complex with 6 μL Alt-R Cas9 enzyme and incubating at room temperature for 10–20 minutes. Two million 518T2 cells were resuspended in 120 μL of SG Cell Line 4D-Nucleofector™ X Kit L solution (Lonza; Cat. #: V4XC-3024). The transfection mix was prepared using 15 μL of RNP complex and 3.6 μL of 100 μM Ultramer ssODN donor encoding the Psmb5 C133W mutation (sequence: 5′-T*G*G GGG TGC AGC GGA TTG CAG CTT CTG GGA GCG GTT GTT GGC TCG ACA GTG GCG AAT CTA TGA GCT TCG CAA TAA GGA ACG CAT CTC GG T*C*G-3′), which was then added to 60 μL of the prepared cell suspension. Nucleofection was performed using the 4D-Nucleofector™ core unit from Lonza. After nucleofection, cells were plated into one well of a 6-well plate in 2 mL of DMEM supplemented with 5% FBS.

After 48 hours, sham-treated cells and cells nucleofected with the Psmb5C133W ssODN donor were plated into 6-well plates at 0.5 million cells per well. Cells were treated with 15 nM bortezomib for 10 days, with media replenished every 3–4 days. After selection, bortezomib-containing media was replaced with fresh media without drug, and surviving cells were allowed to recover and expand for one additional week. Surviving cells were then stained with crystal violet staining solution prepared with 0.5% crystal violet (Sigma-Aldrich; #C6158–50G) in 20% ethanol.

Whole Exome Sequencing Analysis

Whole-exome sequencing of Lurbinectedin-resistant clones was performed by AZENTA Life Sciences using SureSelect Mouse All Exon V1 (mSCLC) and Illumina NovaSeq 6000. The analysis workflow was based on Genome Analysis Toolkit (GATK, v4.6.2.0) best practices28, 29. The quality of sequencing reads was evaluated using FastQC (v0.12.1)30 with aggregated reports generated by MultiQC (v1.30)31, and adapter trimming was performed using Trim Galore (v0.6.10)32. High-quality reads were mapped to the mouse reference genome GRCm39 (Ensembl release 114) using BWA-MEM2 (v2.3)33. GATK MarkDuplicates was used to flag PCR duplicates and GATK BaseRecalibrator was applied to recalibrate base quality scores. Variant calling was restricted to exonic regions defined by the GRCm39 Ensembl release 114 GTF annotation. Post-alignment quality metrics, including duplication rate and on-target rate, were assessed using Qualimap (v2.2.2)34. Somatic variant calling was performed using GATK Mutect2 in paired tumor–normal mode, with the CBC parental cell line serving as the matched normal control. Variant filtering was applied using GATK FilterMutectCalls with default parameters. Variants were annotated using Ensembl Variant Effect Predictor (VEP, v114.2)35 against the Mus musculus GRCm39 database. Secondary filtering was performed in R using the vcfR package (v1.15.0)36; somatic mutations were defined by a tumor allele frequency greater than 0.2 (VAF_mut > 0.25) and a matched normal allele frequency below 0.01 (VAF_norm < 0.01). The complete pipeline was implemented in Nextflow37 and executed on a SLURM high-performance computing cluster.

The accession number for the sequences reported in this paper is SRA: PRJNA1473291.

Mutational Signature Analysis

Somatic single-base substitutions from whole-exome sequencing (WES) data were analyzed using COSMIC SigProfilerAssignment. For lurbinectedin-resistant ultramutator clones, variants were included if they were present in the resistant clone at variant allele fraction (VAF) > 0.20 and absent or present at very low frequency in the parental cell line (VAF < 0.01). Only single-nucleotide substitutions with A/C/G/T reference and alternate alleles were included for SBS analysis. Insertions, deletions, multinucleotide variants, non-primary chromosome calls, and duplicate annotation entries corresponding to the same genomic substitution were excluded. Variants were classified into SBS96 trinucleotide contexts using the mm39/GRCm39 reference genome and analyzed with SigProfilerAssignment using the SBS variant class and WES/exome settings. Signature activities were reported as the number of mutations assigned to each COSMIC SBS signature in each clone.

For comparison with an MMR-deficient-only mutational process, we reanalyzed available WES data from previously generated Compound 27-resistant Msh2-KO clones8. Because these data were generated using Ensembl release 91/GRCm38 coordinates, the genome build was specified as mm10 for SigProfilerAssignment. These clone variants were processed using the same SBS-focused workflow. Briefly, somatic single-nucleotide substitutions were retained if they had VAF ≥ 0.20, alternate read depth ≥ 3, and total depth ≥ 10 in the clone. Repeated transcript-level annotations were collapsed so that each genomic position was counted only once per clone. The resulting mutation files were formatted as COSMIC/SigProfilerAssignment simple text input files with the fields Project, Sample, ID, Genome, mut_type, chrom, pos_start, pos_end, ref, alt, and Type. Signature assignment was then performed using the SBS variant class and WES/exome settings.

Radiation treatment

For UV irradiation, cells growing exponentially on culture dishes were washed with phosphate-buffered saline (PBS) and exposed to UV-C (254 nm) at a rate of 0.4–0.5 J/m2/s to achieve the desired cumulative dose. Fresh culture medium was added to the culture dishes immediately after irradiation. The colonies were fixed and stained with 0.5% crystal violet (20% Methanol in PBS).

Quantification and statistical analysis

Data was analyzed using Prism 11.0 by GraphPad. Fitting for dose response curves were performed by using log(inhibitor) vs. Normalized response -- Variable slope (four parameters). This was used to calculate EC50 of each dose response curve.

Statistical significance of resistant clones under bortezomib selection was evaluated by Pearson’s chi-square test of independence. We quantified clone-positive wells across three bortezomib concentrations (7.71, 8.20, and 8.72 nM), corresponding to 288 wells per mutagenesis condition. The 7.24 nM concentration was excluded from the primary analysis because clone-positive outgrowth was observed in the DMSO control at this dose, consistent with incomplete killing and insufficient selective stringency to distinguish resistant outgrowth from background survival. we observed clone-positive wells in the DMSO condition, suggesting that clones may be observed due to lack of complete killing and not due to mutagenesis. Across the three concentrations analyzed, we identified 3, 11, 10, and 27 clone-positive wells in the DMSO, MMRd, Polεmut, and combined Ultra conditions, respectively. We conducted a chi-square test of independence on the resulting 4×2 contingency table using GraphPad Prism. (χ2(3) = 25.34, p < 0.0001). Post-hoc pairwise chi square tests between conditions was performed with chi-square analysis on 2×2 contingency tables.

Supplementary Material

Supplementary Figures
Bortezomib Resistant Clones_Data
WES Lurbinectedin Clones_Data

ASSOCIATED CONTENT

Supporting Information:

  • Additional experimental details and data.

  • Table with bortezomib-resistant clones and Psmb5 mutations identified

  • Whole-exome sequencing analysis of lurbinectedin-resistant clones.

ACKNOWLEDGEMENTS

The authors thank Florentina Vandiver for technical support. psPAX2 and pMD2.G were a gift from the McFadden Lab. XLone and pCAGE-HyPBase plasmids were a gift from Dr. Frederick Rehfeld. We thank Jeon Lee and Khoi for the support with the bioinformatic analysis.

Funding Sources

D.G.M. was supported by NIH U54CA231649 and NIH UM1CA294119. A.J.D. was supported by a Department of Energy DE-SC0025578, NIH/NINDS R01NS133439, NIH/NCI R01CA276058, and NASA 80NSSC23K1018.

REFERENCES

  • 1.Schenone M; Dancik V; Wagner BK; Clemons PA Target identification and mechanism of action in chemical biology and drug discovery. Nat Chem Biol 2013, 9 (4), 232–240. DOI: 10.1038/nchembio.1199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Wacker SA; Houghtaling BR; Elemento O; Kapoor TM Using transcriptome sequencing to identify mechanisms of drug action and resistance. Nat Chem Biol 2012, 8 (3), 235–237. DOI: 10.1038/nchembio.779. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Han T; Goralski M; Capota E; Padrick SB; Kim J; Xie Y; Nijhawan D The antitumor toxin CD437 is a direct inhibitor of DNA polymerase alpha. Nat Chem Biol 2016, 12 (7), 511–515. DOI: 10.1038/nchembio.2082. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Horn M; Kroef V; Allmeroth K; Schuller N; Miethe S; Peifer M; Penninger JM; Elling U; Denzel MS Unbiased compound-protein interface mapping and prediction of chemoresistance loci through forward genetics in haploid stem cells. Oncotarget 2018, 9 (11), 9838–9851. DOI: 10.18632/oncotarget.24305. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Povedano JM; Liou J; Wei D; Srivatsav A; Kim J; Xie Y; Nijhawan D; McFadden DG Engineering Forward Genetics into Cultured Cancer Cells for Chemical Target Identification. Cell Chem Biol 2019, 26 (9), 1315–1321 e1313. DOI: 10.1016/j.chembiol.2019.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Nguyen TP; Fang M; Kim J; Wang B; Lin E; Khivansara V; Barrows N; Rivera-Cancel G; Goralski M; Cervantes CL; et al. Inducible mismatch repair streamlines forward genetic approaches to target identification of cytotoxic small molecules. Cell Chem Biol 2023, 30 (11), 1453–1467 e1458. DOI: 10.1016/j.chembiol.2023.07.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Povedano JM; Li V; Lake KE; Bai X; Rallabandi R; Kim J; Xie Y; De Brabander JK; McFadden DG TK216 targets microtubules in Ewing sarcoma cells. Cell Chem Biol 2022, 29 (8), 1325–1332 e1324. DOI: 10.1016/j.chembiol.2022.06.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Povedano JM; Rallabandi R; Bai X; Ye X; Liou J; Chen H; Kim J; Xie Y; Posner B; Rice L; et al. A Multipronged Approach Establishes Covalent Modification of beta-Tubulin as the Mode of Action of Benzamide Anti-cancer Toxins. J Med Chem 2020, 63 (22), 14054–14066. DOI: 10.1021/acs.jmedchem.0c01482. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Theodoropoulos PC; Wang W; Budhipramono A; Thompson BM; Madhusudhan N; Mitsche MA; McDonald JG; De Brabander JK; Nijhawan D A Medicinal Chemistry-Driven Approach Identified the Sterol Isomerase EBP as the Molecular Target of TASIN Colorectal Cancer Toxins. J Am Chem Soc 2020, 142 (13), 6128–6138. DOI: 10.1021/jacs.9b13407. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Han T; Goralski M; Gaskill N; Capota E; Kim J; Ting TC; Xie Y; Williams NS; Nijhawan D Anticancer sulfonamides target splicing by inducing RBM39 degradation via recruitment to DCAF15. Science 2017, 356 (6336). DOI: 10.1126/science.aal3755. [DOI] [PubMed] [Google Scholar]
  • 11.Shlien A; Campbell BB; de Borja R; Alexandrov LB; Merico D; Wedge D; Van Loo P; Tarpey PS; Coupland P; Behjati S; et al. Combined hereditary and somatic mutations of replication error repair genes result in rapid onset of ultra-hypermutated cancers. Nat Genet 2015, 47 (3), 257–262. DOI: 10.1038/ng.3202. [DOI] [PubMed] [Google Scholar]
  • 12.Nishimura K; Fukagawa T; Takisawa H; Kakimoto T; Kanemaki M An auxin-based degron system for the rapid depletion of proteins in nonplant cells. Nat Methods 2009, 6 (12), 917–922. DOI: 10.1038/nmeth.1401. [DOI] [PubMed] [Google Scholar]
  • 13.Hills-Muckey K; Martinez MAQ; Stec N; Hebbar S; Saldanha J; Medwig-Kinney TN; Moore FEQ; Ivanova M; Morao A; Ward JD; et al. An engineered, orthogonal auxin analog/AtTIR1(F79G) pairing improves both specificity and efficacy of the auxin degradation system in Caenorhabditis elegans. Genetics 2022, 220 (2). DOI: 10.1093/genetics/iyab174. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Yamada R; Murai K; Uchida N; Takahashi K; Iwasaki R; Tada Y; Kinoshita T; Itami K; Torii KU; Hagihara S A Super Strong Engineered Auxin-TIR1 Pair. Plant Cell Physiol 2018, 59 (8), 1538–1544. DOI: 10.1093/pcp/pcy127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Li HD; Cuevas I; Zhang M; Lu C; Alam MM; Fu YX; You MJ; Akbay EA; Zhang H; Castrillon DH Polymerase-mediated ultramutagenesis in mice produces diverse cancers with high mutational load. J Clin Invest 2018, 128 (9), 4179–4191. DOI: 10.1172/JCI122095. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Li HD; Lu C; Zhang H; Hu Q; Zhang J; Cuevas IC; Sahoo SS; Aguilar M; Maurais EG; Zhang S; et al. A PoleP286R mouse model of endometrial cancer recapitulates high mutational burden and immunotherapy response. JCI Insight 2020, 5 (14). DOI: 10.1172/jci.insight.138829. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Oerlemans R; Franke NE; Assaraf YG; Cloos J; van Zantwijk I; Berkers CR; Scheffer GL; Debipersad K; Vojtekova K; Lemos C; et al. Molecular basis of bortezomib resistance: proteasome subunit beta5 (PSMB5) gene mutation and overexpression of PSMB5 protein. Blood 2008, 112 (6), 2489–2499. DOI: 10.1182/blood-2007-08-104950. [DOI] [PubMed] [Google Scholar]
  • 18.Palombella VJ; Conner EM; Fuseler JW; Destree A; Davis JM; Laroux FS; Wolf RE; Huang J; Brand S; Elliott PJ; et al. Role of the proteasome and NF-kappaB in streptococcal cell wall-induced polyarthritis. Proc Natl Acad Sci U S A 1998, 95 (26), 15671–15676. DOI: 10.1073/pnas.95.26.15671. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Hess GT; Fresard L; Han K; Lee CH; Li A; Cimprich KA; Montgomery SB; Bassik MC Directed evolution using dCas9-targeted somatic hypermutation in mammalian cells. Nat Methods 2016, 13 (12), 1036–1042. DOI: 10.1038/nmeth.4038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Paz-Ares L; Borghaei H; Liu SV; Peters S; Herbst RS; Stencel K; Majem M; Sendur MAN; Czyzewicz G; Caro RB; et al. Efficacy and safety of first-line maintenance therapy with lurbinectedin plus atezolizumab in extensive-stage small-cell lung cancer (IMforte): a randomised, multicentre, open-label, phase 3 trial. Lancet 2025, 405 (10495), 2129–2143. DOI: 10.1016/S0140-6736(25)01011-6. [DOI] [PubMed] [Google Scholar]
  • 21.Pommier Y; Kohlhagen G; Bailly C; Waring M; Mazumder A; Kohn KW DNA sequence- and structure-selective alkylation of guanine N2 in the DNA minor groove by ecteinascidin 743, a potent antitumor compound from the Caribbean tunicate Ecteinascidia turbinata. Biochemistry 1996, 35 (41), 13303–13309. DOI: 10.1021/bi960306b. [DOI] [PubMed] [Google Scholar]
  • 22.Takebayashi Y; Pourquier P; Yoshida A; Kohlhagen G; Pommier Y Poisoning of human DNA topoisomerase I by ecteinascidin 743, an anticancer drug that selectively alkylates DNA in the minor groove. Proc Natl Acad Sci U S A 1999, 96 (13), 7196–7201. DOI: 10.1073/pnas.96.13.7196. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Takebayashi Y; Pourquier P; Zimonjic DB; Nakayama K; Emmert S; Ueda T; Urasaki Y; Kanzaki A; Akiyama SI; Popescu N; et al. Antiproliferative activity of ecteinascidin 743 is dependent upon transcription-coupled nucleotide-excision repair. Nat Med 2001, 7 (8), 961–966. DOI: 10.1038/91008. [DOI] [PubMed] [Google Scholar]
  • 24.Vidal A; Munoz C; Guillen MJ; Moreto J; Puertas S; Martinez-Iniesta M; Figueras A; Padulles L; Garcia-Rodriguez FJ; Berdiel-Acer M; et al. Lurbinectedin (PM01183), a new DNA minor groove binder, inhibits growth of orthotopic primary graft of cisplatin-resistant epithelial ovarian cancer. Clin Cancer Res 2012, 18 (19), 5399–5411. DOI: 10.1158/1078-0432.CCR-12-1513. [DOI] [PubMed] [Google Scholar]
  • 25.Marteijn JA; Lans H; Vermeulen W; Hoeijmakers JH Understanding nucleotide excision repair and its roles in cancer and ageing. Nat Rev Mol Cell Biol 2014, 15 (7), 465–481. DOI: 10.1038/nrm3822. [DOI] [PubMed] [Google Scholar]
  • 26.Allmeroth K; Horn M; Kroef V; Miethe S; Muller RU; Denzel MS Bortezomib resistance mutations in PSMB5 determine response to second-generation proteasome inhibitors in multiple myeloma. Leukemia 2021, 35 (3), 887–892. DOI: 10.1038/s41375-020-0989-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Santamaria Nunez G; Robles CM; Giraudon C; Martinez-Leal JF; Compe E; Coin F; Aviles P; Galmarini CM; Egly JM Lurbinectedin Specifically Triggers the Degradation of Phosphorylated RNA Polymerase II and the Formation of DNA Breaks in Cancer Cells. Mol Cancer Ther 2016, 15 (10), 2399–2412. DOI: 10.1158/1535-7163.MCT-16-0172. [DOI] [PubMed] [Google Scholar]
  • 28.DePristo MA; Banks E; Poplin R; Garimella KV; Maguire JR; Hartl C; Philippakis AA; del Angel G; Rivas MA; Hanna M; et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet 2011, 43 (5), 491–498. DOI: 10.1038/ng.806. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.McKenna A; Hanna M; Banks E; Sivachenko A; Cibulskis K; Kernytsky A; Garimella K; Altshuler D; Gabriel S; Daly M; et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res 2010, 20 (9), 1297–1303. DOI: 10.1101/gr.107524.110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.FastQC: A Quality Control Tool for High Throughput Sequence Data; 2010. http://www.bioinformatics.babraham.ac.uk/projects/fastqc/ (accessed. [Google Scholar]
  • 31.Ewels P; Magnusson M; Lundin S; Kaller M MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics 2016, 32 (19), 3047–3048. DOI: 10.1093/bioinformatics/btw354. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Trim Galore; 2012. https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/ (accessed. [Google Scholar]
  • 33.Vasimuddin M; Misra S; Li H; Aluru S Efficient Architecture-Aware Acceleration of BWA-MEM for Multicore Systems. Int Parall Distrib P 2019, 314–324. DOI: 10.1109/Ipdps.2019.00041. [DOI] [Google Scholar]
  • 34.Garcia-Alcalde F; Okonechnikov K; Carbonell J; Cruz LM; Gotz S; Tarazona S; Dopazo J; Meyer TF; Conesa A Qualimap: evaluating next-generation sequencing alignment data. Bioinformatics 2012, 28 (20), 2678–2679. DOI: 10.1093/bioinformatics/bts503. [DOI] [PubMed] [Google Scholar]
  • 35.McLaren W; Gil L; Hunt SE; Riat HS; Ritchie GR; Thormann A; Flicek P; Cunningham F The Ensembl Variant Effect Predictor. Genome Biol 2016, 17 (1), 122. DOI: 10.1186/s13059-016-0974-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Knaus BJ; Grunwald NJ vcfr: a package to manipulate and visualize variant call format data in R. Mol Ecol Resour 2017, 17 (1), 44–53. DOI: 10.1111/1755-0998.12549. [DOI] [PubMed] [Google Scholar]
  • 37.Di Tommaso P; Chatzou M; Floden EW; Barja PP; Palumbo E; Notredame C Nextflow enables reproducible computational workflows. Nat Biotechnol 2017, 35 (4), 316–319. DOI: 10.1038/nbt.3820. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Figures
NIHMS2209994-supplement-Supplementary_Figures.pdf (4.3MB, pdf)
Bortezomib Resistant Clones_Data
NIHMS2209994-supplement-Bortezomib_Resistant_Clones_Data.xlsx (16.9KB, xlsx)
WES Lurbinectedin Clones_Data
NIHMS2209994-supplement-WES_Lurbinectedin_Clones_Data.xlsx (423.2KB, xlsx)

RESOURCES