News Article | December 12, 2016
Science has always issued medical promissory notes. In the 17th century, Francis Bacon promised that an understanding of the true mechanisms of disease would enable us to extend life almost indefinitely; René Descartes thought that 1,000 years sounded reasonable. But no science has been more optimistic, more based on promises, than medical genetics. Recently, I read an article promising that medical genetics will soon deliver 'a world in which doctors come to their patients and tell them what diseases they are about to have'. Treatments can begin 'before the patient feels even the first symptoms!' So promises 'precision medicine', which aims to make medicine predictive and personalized through detailed knowledge of the patient's genome. The thing is, the article is from 1940. It's a yellowed scrap of newsprint in the Alan Mason Chesney Archives at Johns Hopkins University in Baltimore. The article profiles Madge Thurlow Macklin, a Hopkins-trained physician working at the University of Western Ontario. Macklin's mid-century genetics is not today's genetics. In 1940, genes were made of protein, not DNA. Textbooks stated that we have 48 chromosomes (we have 46). Looking back, we knew almost exactly nothing about the genetic mechanisms of human disease. These genetic promissories echo down the decades with an eerie resonance. In 1912, Harvey Ernest Jordan—who would become dean of the University of Virginia medical school—wrote: 'Medicine is fast becoming a science of the prevention of weakness and morbidity; their permanent not temporary cure, their racial eradication rather than their personal palliation.' (By 'racial' here Jordan simply meant any large, loosely related population.) 'Fast' is relative; 99 years later, in 2011, Leroy Hood wrote: 'Medicine will move from a reactive to a proactive discipline over the next decade.' Cancer is often named as precision medicine's most promising focus. In 2003, Andrew C von Eschenbach, the head of the National Cancer Institute, set a goal of eliminating death and suffering due to cancer—by 2015. On 20 September this year, Microsoft announced an initiative to cure cancer by 2026. Jasmin Fisher, a senior researcher on the project, said: 'If we are able to control and regulate cancer then it becomes like any chronic disease and then the problem is solved.' Such statements bring to mind the old Monty Python sketch, 'How to Do It', in which Eric Idle (in drag) explains how to rid the world of all known diseases. 'Well, first of all become a doctor and discover a marvellous [sic] cure for something,' he announces. 'Then,' he continues, 'when the medical world really starts to take notice of you, you can jolly well tell them what to do and make sure they get everything right so there'll never be diseases any more.' Just a day after Microsoft's announcement, on 21 September, Mark Zuckerberg and his wife, the pediatrician Priscilla Chan, announced that they are giving $3 billion to their Chan Zuckerberg Initiative (CZI). The CZI aims to 'cure all disease in our children's lifetime'. Once again, the Pythons nailed it. With their money, the doctors can discover "marvellous" cures for things. Then, when the medical world really starts to take notice, CZI can jolly well tell them what to do and make sure they get everything right, so there'll never be any diseases any more. Read More: There's One Problem With Mark Zuckerberg's Plan to Solve All Diseases While inflated medical promises are hardly peculiar to molecular medicine, that field does seem particularly prone to breathless rhetoric. You can almost hear K Eric Drexler panting when he writes, in his manifesto Engines of Creation (1986), that protein-based nanomachines 'promise to bring changes as profound as the Industrial Revolution, antibiotics, and nuclear weapons all rolled up in one massive breakthrough'. Bluster, overstatement and aspirations masquerading as hard targets have no single cause. One reason, surely, is the heady sense of impending omnipotence that accompanies major technological and scientific advances. Charles Darwin's theory of evolution by natural selection, the rediscovery of Gregor Mendel's laws of heredity, the cracking of the genetic code, genetic engineering, the Human Genome Project, CRISPR—all were followed by grandiose claims of the imminent total control over life's fundamental processes. Every generation of scientists looks back and shakes its collective head in condescending disbelief at how little the previous generation knew, rarely stopping to reflect that the next generation will do the same. In our particular moment, biology is the king, and the perennial desire for simple solutions to complex problems leads people back time and again to biological determinism: it's all in your genes. It's all in your neurons. This new discovery changes everything. The possibility of vast profits in biotech also contribute to the propensity for hype. Buy on the rumors, sell on the news. Nowhere is that more true than in biotech and infotech. As the molecular biologist James Watson—no stranger to hype himself—wrote in his memoir Avoid Boring People (2007): 'Nothing attracts money like the quest for a cure for a terrible disease.' Finally, the researchers and their funders vie for attention from a news media that is itself constantly competing for an overstimulated and numb audience. Overcoming habituation and making a splash requires ever-bigger jolts of hyperbole. It's time to push back. One way is to hold scientists, philanthropists and the press accountable. In 2014, Jonathan Eisen, professor at the Genome Center at the University of California, Davis, compiled a lengthy list of articles on the hype surrounding the genome project—many of them either complaining of promise fatigue or pricking the bubble of inflated expectations. We can and should continue writing, collecting and sharing such pieces. Fund science liberally, but reward knowledge more than market value. Encourage science literacy, not just cheerleading. And teach skepticism of technology, medicine and the media. Now, that might sound like a pipe dream itself. Nevertheless, any progress in this direction will yield results. Science does lead to better understanding, and new knowledge will continue to yield new drugs and other therapies. But better understanding also means an appreciation for how complex nature is. The progress of science is the steady realization of how little we actually know. The more we, the public, understand that about science, the more we will see through the hype. And the more we see through the hype, the more medicine will serve its stakeholders, not its stockholders. We will get better healthcare. Nathaniel Comfort is the Baruch Blumberg Chair of Astrobiology at NASA/Library of Congress and professor of history of medicine at Johns Hopkins University in Baltimore. His latest book is The Science of Human Perfection: How Genes Became the Heart of American Medicine (2012). This article was originally published at Aeon and has been republished under Creative Commons.
News Article | October 5, 2016
No statistical methods were used to pre-evaluate the sample size in this study. The experiments (including animal experiments) were not randomized. The investigators were not blinded to experiments. No samples/data were excluded except any obviously unhealthy xenografted mice. H1299, U2OS, MCF7, H460 and HCT116 cell lines were cultured in DMEM supplemented with 10% (vol/vol) FBS. The SU-DHL-5 cell line was cultured in IMDM supplemented with 10% (vol/vol) FBS. MEFs were cultured in DMEM supplemented with 10% (vol/vol) heat-inactivated FBS. All the cell lines were obtained from ATCC and have been proven to be negative for mycoplasma contamination. No cell lines used in this work were listed in the ICLAC database. The cell lines were freshly thawed from the purchased seed cells and were cultured for no more than 2 months. The morphology of cell lines was checked every week and compared with the ATCC cell line image to avoid cross-contamination or misuse of cell lines. SET stable knockdown cells were generated by lentivirus-based infection of shRNA. SET cDNA was purchased from Addgene (Plasmid number 24998) and the full-length cDNA or the various fragments were sub-cloned into pWG-F-HA, pCMV-Myc or PGEX-2TL vectors. Each p53 plasmid was generated by sub-cloning human p53 cDNA (including full-length or various fragments) into pWG-F-HA, pcDNA3.1 or PGEX-2TL vectors. The point-mutation constructs (including p53-KR and -KQ) were generated by using a site-directed mutagenesis Kit (Stratagene, 200521). Introduction of the expressing construct and siRNA transfection were performed by Lipofectamine 2000 (Invitrogen, 11668-019) according to the manufacturer’s protocol. To transfer oligos into SU-DHL-5 cells, we used electroporation following the manufacturer’s protocol (Lonza PBC3-00675). The DNA damage inducer doxorubicin was used at 1 μM for 24 h. The proteasome inhibitor epoxomicin was used at 100 nM for 6 h. Cells were treated with TSA (1 μM) and nicotinamide (5 mM) for 6 h to inhibit HDAC activity in the assays in which p53 acetylation needed to be maintained. Ad–GFP and Ad–Cre–GFP viruses were purchased from Vector Biolabs (Catalogue numbers 1761 and 1710). To generate the knock-in mice, W4/129S6 mouse embryonic stem (ES) cells (Taconic) were electroporated with a targeting vector containing homologous regions flanking the mouse p53 exon 11, in which all 7 lysines were mutated to glutamines (p53KQ allele). A neomycin-resistance gene cassette flanked by two LoxP sites (LNL) was inserted into intron 10 to allow selection of targeted ES cell clones with G418. ES cell clones were screened by Southern blotting with EcoRI-digested genomic DNA, using a probe generated from PCR amplification in the region outside the homologous region in the targeting vector. The correctly targeted ES cell clones containing the K-to-Q mutations were injected into C57BL/6 blastocysts, which were then implanted into pseudopregnant females to generate chimaeras. Germ-line transmission was accomplished by breeding chimaeras with C57BL/6 mice. Subsequently, mice containing the targeted allele were bred with Rosa26-Cre mice to remove the LNL cassette and to generate mice with only the K-to-Q mutations. To confirm the mutations inserted in p53+/KQ mice, we sequenced p53 cDNA derived from mRNA isolated from p53+/KQ spleen. All seven K-to-Q mutations were confirmed and no additional mutations were found. The offspring were genotyped by PCR using the following primer set, forward: 5′-GGGAGGATAAACTGATTCTCAGA-3′, reverse: 5′-GATGGCTTCTACTATGGGTAGGGAT-3′. To generate a Set conditional knockout mouse, exon 2 of the Set gene was floxed and deletion of exon 2 resulted in a frameshift and the truncation of the C-terminal domain. The targeting vector of Set contained 10 kb genomic DNA spanning exon 2; a neomycin-resistance gene cassette and loxP sites were inserted flanking exon 2. To increase targeting frequency, a diphtheria toxin A cassette was inserted at the 3′ end of the targeting vector to reduce random integration of the modified Set genomic DNA. A new BglII restriction site was also inserted to facilitate Southern blot screening. Of the 200 mouse ES cell clones screened, eight were identified to have integrated the floxed exon 2 by Southern blot using a 5′ probe, which detects a 14-kb band for the wild-type allele and an 11-kb band for the floxed exon 2 allele (Setflox). Two of the clones were then injected into blastocysts to generate Set chimaera mice and they were bred to produce germ-line transmission of the floxed exon 2 allele. Setflox/+ mice were intercrossed to generate Set homozygous conditional knockout mice (Setflox/flox). Maintenance and experimental procedures of mice were approved by the Institutional Animal Care and Use Committee (IACUC) of Columbia University. For the in vitro peptide binding assay: equal amounts of each synthesized biotin-conjugated peptide (made as column or as batch) were incubated with highly concentrated HeLa nuclear extract (NE) or purified proteins for 1 h or overnight at 4 °C. After washing with BC100 buffer (20 mM Tris-HCl pH 7.9, 100 mM NaCl, 10% glycerol, 0.2 mM EDTA, 0.1% triton X-100) three times, the binding components were eluted in high-salt buffer (20 mM Tris-HCl pH 7.9, 1,000 mM NaCl, 1% DOC, 10% glycerol, 0.2 mM EDTA, 0.1% triton X-100) or by boiling with 1 × Laemmli buffer for further analysis. For the in vitro GST-fusion protein binding assay: Escherichia coli containing GST or GST-fusion protein expressing constructs were grown in a shaking incubator at 37 °C until the OD was about 0.6. Next 0.1 mM IPTG was added and the E. coli were incubated at 25 °C for 4 h or overnight, to induce GST or GST-fusion protein expression. After purification by GST·Bind Resin (Novagen, 70541), equal amounts of immobilized GST or GST-fusion proteins were incubated with other purified proteins for 1 h at 4 °C, followed by washing with BC100 buffer three times. The binding components were eluted by boiling with 1 × Laemmli buffer and were analysed by western blot. Whole cellular extracts (WCE) were prepared in BC100 buffer with sonication. Nuclear extract (NE) was prepared by sequentially lysing cells with HB buffer (20 mM Tris-HCl pH 7.9, 10 mM KCl, 1.5 mM MgCl , 1 mM PMSF, 1 × protease inhibitor (Sigma)) for the cytosolic fraction and BC400 buffer (20 mM Tris-HCl pH 7.9, 400 mM NaCl, 10% Glycerol, 0.2 mM EDTA, 0.5% triton X-100, 1 mM PMSF, 1 × protease inhibitor) for nuclear fraction. The salt concentration of NE was adjusted to 100 mM. 2 μg of the indicated antibody (or 20 μl Flag M2 Affinity Gel (Sigma, A2220)) was added into WCE or NE and incubated overnight at 4 °C, followed by addition of 20 μl protein A/G agarose (Santa Cruz, sc-2003; only for IP with unconjugated antibodies mentioned above) for 2 h. After washing with BC100 buffer three times, the binding components were eluted using Flag peptide (Sigma, F3290), 0.1% trifluoroacetic acid (TFA, Sigma, 302031) or by boiling with 1 × Laemmli buffer, and were analysed by western blot. For preparation of Ub-p53: H1299 cells were co-transfected with p53, MDM2 and 6 × HA-Ub (human) expressing plasmids for 48 h. The cells were lysed with Flag lysis buffer (50 mM Tris-HCl pH 7.9, 137 mM NaCl, 10 mM NaF, 1 mM Na VO , 10% glycerol, 0.5 mM EDTA, 1% triton X-100, 0.2% sarkosyl (sodium lauroyl sarcosinate), 0.5 mM DTT, 1 mM PMSF, 1 × protease inhibitor) and total Ub-conjugated proteins were purified by anti-HA-agarose (Sigma, A2095) and eluted by 1 × HA peptide (Sigma I2149). For the preparation of Sumo-p53 or Nedd-p53: H1299 cells were co-transfected with p53, MDM2 (only for Nedd-p53 preparation) and 6 × His-HA-Sumo1 (human) or 6 × His-HA-Nedd8 (human) expressing plasmids for 48 h. The cells were lysed with guanidine lysis buffer (6 M guanidin-HCl, 0.1 M Na HPO , 6.8 mM NaH PO , 10 mM Tris-HCl pH 8.0, 0.2% triton-X100, freshly supplemented with 10 mM β-mercaptoethanol and 5 mM imidazole) with mild sonication. After overnight pull-down by Ni+-NTA agarose (Qiagen 30230), the binding fractions were sequentially washed with guanidine lysis buffer, urea buffer I (8 M urea, 0.1 M Na HPO , 6.8 mM NaH PO , 10 mM Tris-HCl pH 8.0, 0.2% triton-X100, freshly supplemented with 10 mM β-mercaptoethanol and 5 mM imidazole) and urea buffer II (8 M urea, 18 mM Na HPO , 80 mM NaH PO , 10 mM Tris-HCl pH 6.3, 0.2% triton-X100, freshly supplemented with 10 mM β-mercaptoethanol and 5 mM imidazole). Precipitates were eluted in elution buffer (0.5 M imidazole, 0.125 M DTT). All purified proteins were dialysed against BC100 buffer before use in the subsequent pull-down assay. After the pull-down assay, the interaction between SET and each p53-conjugate was detected by western blot with anti-p53 (DO-1) antibody. The protein complex was separated by SDS–PAGE and stained with GelCode Blue reagent (Pierce, 24592). The visible band was cut and digested with trypsin and then subjected to liquid chromatography (LC)-MS/MS analysis. A firefly reporter (p21-Luci reporter) and a Renilla control reporter were co-transfected with indicated constructs in H1299 cells for 48 h and the relative luciferase activity was measured by dual-luciferase assay protocol (Promega, E1910). Highly purified p53 or SET was incubated with a 32P-labelled probe (160 bp) containing the p53-binding element of the p21 promoter in 1× binding buffer (10 mM HEPES, pH 7.6, 40 mM NaCl, 50 μM EDTA, 6.25% glycerol, 1 mM MgCl , 1 mM spermidine, 1 mM DTT, 50 ng μl−1 BSA, 5 ng μl−1 sheared single strand salmon DNA) for 20 min at room temperature (RT). For the super-shift assay, α-p53 or α-SET antibody was pre-incubated with purified p53 and SET in the reaction system without probe for 30 min at RT and then the probe was added for a further 20 min. The complex was analysed by 4% Tris-Borate-EDTA buffer–polyacrylamide gel electrophoresis (TBE–PAGE) and visualized by autoradiography. The probe was obtained by PCR, labelled by T4 kinase (NEB, M0201S) and purified by Bio-Spin column (Bio-Rad, 732-6223). Cells were fixed with 1% formaldehyde for 10 min at room temperature and lysed with ChIP lysis buffer (50 mM Tris-HCl pH 8.0, 5 mM EDTA, 1% SDS, 1× protease inhibitor) for 10 min at 4 °C. After sonication, the lysates were centrifuged, and the supernatants were collected and pre-cleaned by salmon sperm DNA saturated protein A agarose (Millipore, 16-157) in dilution buffer (20 mM Tris-HCl pH 8.0, 2 mM EDTA, 150 mM NaCl, 1% triton X-100, 1× protease inhibitor) for 1 h at 4 °C. The pre-cleaned lysates were aliquoted equally and incubated with indicated antibodies overnight at 4 °C. Saturated protein A agarose was added into each sample and incubated for 2 h at 4 °C. The agarose was washed with TSE I (20 mM Tris-HCl pH 8.0, 2 mM EDTA, 150 mM NaCl, 0.1% SDS, 1% triton X-100), TSE II (20 mM Tris-HCl pH 8.0, 2 mM EDTA, 500 mM NaCl, 0.1% SDS, 1% triton X-100), buffer III (10 mM Tris-HCl pH 8.0, 1 mM EDTA, 0.25 M LiCl, 1% DOC, 1% NP40), and buffer TE (10 mM Tris-HCl pH 8.0, 1 mM EDTA), sequentially. The binding components were eluted in 1% SDS and 0.1 M NaHCO and reverse cross-linkage was performed at 65 °C for at least 6 h. DNA was extracted using the PCR purification Kit (Qiagen, 28106). Real-time PCR was performed to detect relative enrichment of each protein or modification on indicated genes. Approximately 105 MEFs or U2OS cells, as indicated in each figure, were seeded into 6-well plates with three replicates. Their cell growth was monitored on consecutive days, as indicated, by using the Countess automated cell counter (Invitrogen) or by staining with 0.1% crystal violet. For quantitative analysis of the crystal violet staining, the crystal violet was extracted from cells using 10% acetic acid and the relative cell number was measured by detecting the absorbance at 590 nm. 106 HCT116-derived cells, as indicated in each figure, were mixed with Matrigel (Corning, 354248) in a 1:1 ratio in a total volume of 200 μl. The cell–matrix complex was subcutaneously injected into nude mice (NU/NU; 8 weeks old; female; strain 088; Charles River). After 3 weeks, the mice were killed and weight of the tumours was measured. The experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Columbia University. None of the experiments were exceeded the limit for tumour burden (10% of total bodyweight or 2 cm in diameter). Total RNA was extracted by TRIzol (Invitrogen, 15596-026) and precipitated in ethanol. 1 μg of total RNA was reverse transcribed into cDNA using the SuperScript III First-Strand Synthesis SuperMix (Invitrogen, 11752-50). The relative expression of each target was measured by qPCR and the data were normalized by the relative expression of GAPDH or ActB. FFPE sections of mouse brain tissue samples were stained with indicated antibodies and visualized by DAB exposure. The Flag-tagged p53 or SET construct was transfected into H1299 cells for 48 h and the cells were lysed in Flag lysis buffer. After centrifugation, the Flag M2 Affinity Gel was added to supernatant and incubated for 1h at 4 °C. After washing with Flag lysis buffer six times, the purified proteins were eluted with Flag peptide. For purification of acetylated p53, the construct CBP was co-transfected with the p53 vector for 48 h. TSA and nicotinamide were added into the medium for the last 6 h and the cells were harvested in Flag lysis buffer supplemented with TSA and nicotinamide. The C-terminal unacetylated p53 was removed by p53-PAb421 antibody and then the acetylated p53 was purified as described above. 0.5 μg recombinant H3 was incubated with 20 ng purified p300 in 1× HAT buffer (50 mM Tris-HCl, pH 7.9; 1 mM DTT; 10 mM sodium butyrate, 10% glycerol) containing 0.1 mM Ac-CoA for 30 min at 30 °C. After the reaction, the products were assayed by western blot with indicated antibodies. To measure the effect of SET on p300-mediated H3 acetylation, H3 and purified SET (1 μg) were pre-incubated in 1× HAT buffer for 20 min at room temperature before addition of the other components (p300 and Ac-CoA) for the subsequent in vitro acetylation assay. Cells were transfected with constructs expressing Cas9-D10A (Nickase) and control sgRNAs or sgRNAs targeting p53 exon3 (Santa Cruz: sc-437281 for control; sc-416469-NIC for targeting of p53). After 48 h of transfection, cells were suspended, diluted and re-seeded to ensure single clone formation. More than 30 clones were picked up and the expression of p53 in each single clone was evaluated by western blot with both α-p53 (DO-1) and α-p53 (FL-393) antibodies. Further verification of positive clones was done by sequencing the genomic DNA to make sure that the functional genomic editing occured (insertion or deletion-mediated frame-shift of the p53 open reading frame (ORF)). Two (U2OS) or three (HCT116) clones were finally selected for subsequent experiments. The p53 knockout-mediated effect was verified to be reproducible in these independent clones. The targeting sequences of p53 loci for the sgRNAs were: 1) TTGCCGTCCCAAGCAATGGA; 2) CCCCGGACGATATTGAACAA. U2OS (CRISPR Ctr or CRISPR p53-KO) cells were transfected with control siRNA or SET-specific siRNA (three oligos) for 4 days. Each sample group had at least two biological replicates. Total RNA was prepared using TRIzol (Invitrogen, 15596-026). The RNA quality was evaluated by Bioanalyzer (Agilent) and confirmed that the RIN > 8. Before performing RNA-seq analysis, a small aliquot of each sample was analysed by RT–qPCR to confirm SET knockdown efficiency. RNA-seq analysis was performed at the Columbia Genome Center. Specifically, from total RNA samples, mRNAs were enriched by poly-A pull-down and then processed for library preparation by using the Illumina TruSeq RNA prep kit (Illumina RS-122-2001). Libraries were then sequenced using the Illumina HiSeq2000. Samples were multiplexed in each lane and yielded targeted number of single-end 100-bp reads for each sample. RTA (Illumina) was used for base calling and bcl2fastq (version 1.8.4) was used for converting BCL to fastq format, coupled with adaptor trimming. Reads were mapped to a reference genome (Human: NCBI/build37.2) using TopHat (version 2.0.4). Relative abundance of genes and splice isoforms were determined using Cufflinks (version 2.0.2) using the default settings. Differentially expressed genes were tested under various conditions using DEseq, an R package based on a negative binomial distribution that models the number reads from RNA-seq experiments and tests for differential expression. To further analyse the differentially expressed genes in a more reliable interval, the following filter strategies were applied: 1) the average of FPKM (Fragments per kilobase of transcript per million mapped reads) in either sample group exceeded 0.1; 2) the fold change between the CRISPR Ctr/si-Ctr group and the CRISPR Ctr/si-SET group exceeded 2; 3) the P value between the CRISPR Ctr/si-Ctr group and the CRISPR Ctr/si-SET group < 0.01. To retrieve potential p53 target genes which were repressed by SET in a p53-dependent manner, we searched the filtered RNA-seq results using the following strategies: 1) the expression level in the CRISPR Ctr/si-SET group was at least 2-fold higher than that in the CRISPR Ctr/si-Ctr group; 2) the expression level in the CRISPR Ctr/si-SET group was at least 2-fold higher than that in the CRISPR p53-KO/si-SET group. The filtered genes which were also verified as p53 target genes from the literature were collected and presented as a heatmap. For the discovery of acidic domains in the human proteome: our motif-finding algorithm initially searched for sequence motifs with a minimum acidic composition of 76% using a sliding window of 36 residues, as dictated by experimental results. Motifs found to be partially overlapping were merged into single motifs. Flanking non-acidic residues were subsequently cropped-out from the final motif. Motif discovery was carried out using the UniProt database, which contains 20,187 canonical human proteins, that have been manually annotated and reviewed. For prediction of proteins that bound acidic domain-containing proteins and were regulated by acetylation: we identified proteins that can potentially bind long acidic domains in a similar way to p53: using a K-rich region whose binding properties can be regulated by acetylation. We used the training set assembled in SSPKA, which combines lysine acetylation annotations from multiple resources obtained either experimentally or in the scientific literature. This dataset individually lists all annotated acetylation sites for a given protein. We generated acetylation motifs with multiple acetylation sites by clustering those sites found to within a maximum distance of 11 residues in sequence. Following this, we searched for acetylation motifs with five or more lysines where at least three of them are annotated as acetylation sites. Results are shown as means ± s.d. Statistical significance was determined by using a two-tailed, unpaired Student t-test in all figures except those described below. In Fig. 1g, significance was determined by one-way ANOVA with a Bonferroni post hoc test. In Fig. 2d and g and Extended Data Figs 2c, 3b, d, 4f and 7h, statistical significance was measured by two-way ANOVA with a Bonferroni post hoc test. All statistical analysis was performed using GraphPad Prism software. P < 0.05 was denoted as statistically significant.
News Article | December 8, 2016
News Article | September 1, 2016
Home > Press > New optical material offers unprecedented control of light and thermal radiation Abstract: Columbia Engineers discover that samarium nickelate shows promise for active photonic devices - SmNiO3 could potentially transform optoelectronic technologies, including smart windows, infrared camouflage, and optical communications. A team led by Nanfang Yu, assistant professor of applied physics at Columbia Engineering, has discovered a new phase-transition optical material and demonstrated novel devices that dynamically control light over a much broader wavelength range and with larger modulation amplitude than what has currently been possible. The team, including researchers from Purdue, Harvard, Drexel, and Brookhaven National Laboratory, found that samarium nickelate (SmNiO3) can be electrically tuned continuously between a transparent and an opaque state over an unprecedented broad range of spectrum from the blue in the visible (wavelength of 400 nm) to the thermal radiation spectrum in the mid-infrared (wavelength of a few tens of micrometers). The study, which is the first investigation of the optical properties of SmNiO3 and the first demonstration of the material in photonic device applications, is published online today in Advanced Materials. "The performance of SmNiO3 is record-breaking in terms of the magnitude and wavelength range of optical tuning," Yu says. "There is hardly any other material that offers such a combination of properties that are highly desirable for optoelectronic devices. The reversible tuning between the transparent and opaque states is based on electron doping at room temperature, and potentially very fast, which opens up a wide range of exciting applications, such as 'smart windows' for dynamic and complete control of sunlight, variable thermal emissivity coatings for infrared camouflage and radiative temperature control, optical modulators, and optical memory devices." Some of the potential new functions include using SmNiO3's capability in controlling thermal radiation to build "intelligent" coatings for infrared camouflage and thermoregulation. These coatings could make people and vehicles, for example, appear much colder than they actually are and thus indiscernible under a thermal camera at night. The coating could help reduce the large temperature gradients on a satellite by adjusting the relative thermal radiation from its bright and dark side with respect to the sun and thereby prolong the lifetime of the satellite. Because this phase-transition material can potentially switch between the transparent and opaque states with high speed, it may be used in modulators for free-space optical communication and optical radar and in optical memory devices. Researchers have long been trying to build active optical devices that can dynamically control light. These include Boeing 787 Dreamliner's "smart windows," which control (but not completely) the transmission of sunlight, rewritable DVD discs on which we can use a laser beam to write and erase data, and high-data-rate, long-distance fiber optic communications systems where information is "written" into light beams by optical modulators. Active optical devices are not more common in everyday life, however, because it has been so difficult to find advanced actively tunable optical materials, and to design proper device architectures that amplify the effects of such tunable materials. When Shriram Ramanathan, associate professor of materials science at Harvard, discovered SmNiO3's giant tunable electric resistivity at room temperature, Yu took note. The two met at the IEEE Photonics Conference in 2013 and decided to collaborate. Yu and his students, working with Ramanathan, who is a co-author of this paper, conducted initial optical studies of the phase-transition material, integrated the material into nanostructured designer optical interfaces--"metasurfaces"--and created prototype active optoelectronic devices, including optical modulators that control a beam of light, and variable emissivity coatings that control the efficiency of thermal radiation. "SmNiO3 is really an unusual material," says Zhaoyi Li, the paper's lead author and Yu's PhD student, "because it becomes electrically more insulating and optically more transparent as it is doped with more electrons--this is just the opposite of common materials such as semiconductors." It turns out that doped electrons "lock" into pairs with the electrons initially in the material, a quantum mechanical phenomenon called "strong electron correlation," and this effect makes these electrons unavailable to conduct electric current and absorbing light. So, after electron doping, SmNiO3 thin films that were originally opaque suddenly allow more than 70 percent of visible light and infrared radiation to transmit through. "One of our biggest challenges," Zhaoyi adds, "was to integrate SmNiO3 into optical devices. To address this challenge, we developed special nanofabrication techniques to pattern metasurface structures on SmNiO3 thin films. In addition, we carefully chose the device architecture and materials to ensure that the devices can sustain high temperature and pressure that are required in the fabrication process to activate SmNiO3." Yu and his collaborators plan next to run a systematic study to understand the basic science of the phase transition of SmNiO3 and to explore its technological applications. The team will investigate the intrinsic speed of phase transition and the number of phase-transition cycles the material can endure before it breaks down. They will also work on addressing technological problems, including synthesizing ultra-thin and smooth films of the material and developing nanofabrication techniques to integrate the material into novel flat optical devices. "This work is one crucial step towards realizing the major goal of my research lab, which is to make an optical interface a functional optical device," Yu notes. "We envision replacing bulky optical devices and components with 'flat optics' by utilizing strong interactions between light and two-dimensional structured materials to control light at will. The discovery of this phase-transition material and the successful integration of it into a flat device architecture are a major leap forward to realizing active flat optical devices not only with enhanced performance from the devices we are using today, but with completely new functionalities." Yu's team included Ramanathan, his Harvard PhD student You Zhou, and his Purdue postdoctoral fellow Zhen Zhang, who synthesized the phase-transition material and did some of the phase transition experiments (this work began at Harvard and continued when Ramanathan moved to Purdue); Drexel University Materials Science Professor Christopher Li, PhD student Hao Qi, and research scientist Qiwei Pan, who helped make solid-state devices by integrating SmNiO3 with novel solid polymer electrolytes; and Brookhaven National Laboratory staff scientists Ming Lu and Aaron Stein, who helped device nanofabrication. Yuan Yang, Assistant Professor of Materials Science and Engineering in the Department of Applied Physics and Applied Mathematics at Columbia Engineering, was consulted during the progress of this research. ### The study was funded by DARPA YFA (Defense Advanced Research Projects Agency Young Faculty Award), ONR YIP (Office of Naval Research Young Investigator Program), AFOSR MURI (Air Force Office of Scientific Research Multidisciplinary University Research Initiative) on metasurfaces, Army Research Office, and NSF EPMD (Electronics, Photonics, and Magnetic Devices) program. FUNDING: The work was supported by Defense Advanced Research Projects Agency Young Faculty Award (Grant No.D15AP00111), Office of Naval Research Young Investigator Award program (Grant No. N00014-16-1-2442), Air Force Office of Scientific Research (Grant No. FA9550-14-1-0389 through a Multidisciplinary University Research Initiative program, and Grant No. FA9550-12-1-0189), National Science Foundation (Grant No. ECCS-1307948), and Army Research Office (Grant Nos.W911NF-16-1-0042 and W911NF-14-1-0669). Research was carried out in part at the Center for Functional Nanomaterials, Brookhaven National Laboratory, which was supported by the U.S. Department of Energy, Office of Basic Energy Sciences, under Contract No. DE-SC0012704. The authors acknowledge helpful discussions with Yuan Yang, Assistant Professor of Materials Science and Engineering in the Department of Applied Physics and Applied Mathematics, Columbia Engineering. About Columbia University School of Engineering and Applied Science Columbia Engineering is one of the top engineering schools in the U.S. and one of the oldest in the nation. Based in New York City, the School offers programs to both undergraduate and graduate students who undertake a course of study leading to the bachelor's, master's, or doctoral degree in engineering and applied science. Columbia Engineering's nine departments offers 16 majors and more than 30 minors in engineering and the liberal arts, including an interdisciplinary minor in entrepreneurship with Columbia Business School. With facilities specifically designed and equipped to meet the laboratory and research needs of faculty and students, Columbia Engineering is home to a broad array of basic and advanced research installations, from the Columbia Nano Initiative and Data Science Institute to the Columbia Genome Center. These interdisciplinary centers in science and engineering, big data, nanoscience, and genomic research are leading the way in their respective fields while our engineers and scientists collaborate across the University to solve theoretical and practical problems in many other significant areas. For more information, please click If you have a comment, please us. Issuers of news releases, not 7th Wave, Inc. or Nanotechnology Now, are solely responsible for the accuracy of the content.
News Article | February 14, 2017
Scientists have discovered a new “mastermind fusion gene” may be associated with a rare cancer-causing tumor – pheochromocytomas (“pheo”) and paragangliomas, according to a study published Feb. 13 in Cancer Cell, by researchers at the Uniformed Services University (USU) and the National Cancer Institutes’ The Cancer Genome Atlas. This breakthrough discovery could lead to more precise treatment as well as a better understanding of cancer itself. These adrenal gland tumors are often benign, but they can become malignant, and in some cases lead to life-threatening hypertension, arrhythmia, and stroke, but it’s not clear which tumors will become metastatic because of the disease’s rarity and complex biology. Therefore, patients with the metastatic disease have few treatment options and poor prognosis. To help detect genetic mutations and better understand this disease, a group of researchers at USU and the nationwide Cancer Genome Atlas Research Network examined 173 tumors, performing six genomic tests, such as DNA and RNA sequencing. The researchers found what they refer to as the mastermind fusion gene – the first fusion gene associated with this type of tumor. This hybrid gene forms from two previously separate genes and only occurs in a new subtype of this disease. The researchers suggest this disrupts the normal biology of the cell and thus producing tumor cells. The researchers believe this mastermind fusion gene will help describe for some patients why the tumor has developed, and better predict patient outcome. The fusion gene may also lead to future targeted therapy and have implications for other cancers. Additionally, the researchers found 18 “driver” genes in this type of tumor, meaning there are 18 different ways this tumor could become cancerous. This is an unusually large amount of drivers, not typical for many other tumor types, according the study’s senior author Dr. Matthew Wilkerson, associate professor and Bioinformatics Director of The American Genome Center and the Collaborative Health Initiative Research Program at USU. This finding allowed their team to classify tumors into four major molecular subtypes, which could also lead to developing new therapies. “For patients who have this diagnosis, surrounded by its uncertainties, this new discovery sheds light on the disease. We think these results will ultimately lead to individuals and their families having a better understanding of their prognosis and more precise treatment,” Wilkerson said. The paper’s co-senior authors are Dr. Katherine Nathanson, a professor in the division of Translational Medicine and Human Genetics at the University of Pennsylvania’s Abramson Cancer Center, and Dr. Karel Pacak, chief of the section on Medical Neuroendocrinology at the Eunice Kennedy Shriver National Institute of Child Health and Human Development at the National Institutes of Health.
News Article | February 28, 2017
SHANGHAI, CAMBRIDGE, Mass. and REYKJAVIK, Iceland, Feb. 28, 2017 /PRNewswire/ -- WuXi NextCODE, a WuXi AppTec group company and the contract genomics organization enabling precision medicine worldwide, today announced that the company's chief operating officer Hannes Smarason has been appointed as the company's chief executive officer, and WuXi AppTec senior vice presidents John Long and Alex Fowkes have been named chief financial officer and chief operating officer, respectively. "WuXi NextCODE is executing on its vision to enable anyone to use the genome to advance health and wellness worldwide," said Dr Ge Li, founder and chairman of WuXi AppTec group and chairman of WuXi NextCODE. "It is also positioning itself as the data management platform at the heart of the genomics revolution and what may become the world's largest data ecosystem. Hannes, John and Alex have the breadth of vision and proven executive capabilities to carry out this strategy and advance the company to the next level." "This is an exciting time at WuXi NextCODE, as we work on leading projects in every facet of genomics and advance a global standard for the way genomic data is organized, mined and shared," said Mr Smarason. "John and Alex's leadership and experience in life sciences finance and operations, and in the recent independent listing of other WuXi AppTec companies, will be invaluable as we grow our business and forge our own strategic path. We are very pleased to have them join our team and offer them a warm welcome to WuXi NextCODE." Hannes Smarason co-founded NextCODE Health in 2013 as a spinout from deCODE genetics. He oversaw NextCODE's acquisition by WuXi AppTec in 2015 and its merger with the WuXi Genome Center to create WuXi NextCODE. He served as COO until January 2017 and as CEO will lead the company's overall business and corporate strategy. John Long has served as WuXi AppTec's senior vice president of finance since 2013. He was actively involved in WuXi AppTec's privatization from NYSE in 2015 and played important roles in WuXi AppTec's subsequent corporate restructuring, supporting WuXi Biologics's recent IPO filing process in Hong Kong as well as private placements in the China capital market. John has over twenty years' experience in financial management, investment and operations in the US, China and Singapore, and prior to joining WuXi AppTec served in senior roles at Willis Group, Tyco International and Lucent Technologies. In addition to finance and operations leadership, he will provide WuXi NextCODE with global expertise in governance, reporting, strategic planning, treasury and tax. John holds a bachelor's degree in economics from the University of International Business and Economics in Beijing and received his MBA from Wharton School of Business at University of Pennsylvania. Alex Fowkes has twenty years experience in the life science industry in operations, business development, strategy and legal roles. He joined WuXi AppTec in 2012 initially to lead corporate development and then most recently serving as senior vice president of commercial operations. Prior to WuXi Alex served in a variety roles for Pfizer in the US, UK, Australia and China over 14 years. About WuXi NextCODE WuXi NextCODE is a fully integrated global contract genomics organization. With offices in Shanghai; Kendall Square in Cambridge, Massachusetts; and Reykjavik, Iceland, we offer comprehensive services that enable population, precision medicine, diagnostics and wellness initiatives and enterprises to use the genome to improve health around the world. Our capabilities span study design, sequencing, secondary analysis, storage, and interpretation and scalable analytics – all backed by the most proven and widely used technology for organizing, mining and sharing genome sequence data. We are also applying the same capabilities to advance a growing range of sequence-based tests and scans in China. WuXi NextCODE is a WuXi AppTec Group company. Visit us on the web at wuxinextcode.com.
News Article | March 2, 2017
Humanity may soon generate more data than hard drives or magnetic tape can handle, a problem that has scientists turning to nature's age-old solution for information-storage--DNA. In a new study in Science, a pair of researchers at Columbia University and the New York Genome Center (NYGC) show that an algorithm designed for streaming video on a cellphone can unlock DNA's nearly full storage potential by squeezing more information into its four base nucleotides. They demonstrate that this technology is also extremely reliable. DNA is an ideal storage medium because it's ultra-compact and can last hundreds of thousands of years if kept in a cool, dry place, as demonstrated by the recent recovery of DNA from the bones of a 430,000-year-old human ancestor found in a cave in Spain. "DNA won't degrade over time like cassette tapes and CDs, and it won't become obsolete--if it does, we have bigger problems," said study coauthor Yaniv Erlich, a computer science professor at Columbia Engineering, a member of Columbia's Data Science Institute, and a core member of the NYGC. Erlich and his colleague Dina Zielinski, an associate scientist at NYGC, chose six files to encode, or write, into DNA: a full computer operating system, an 1895 French film, "Arrival of a train at La Ciotat," a $50 Amazon gift card, a computer virus, a Pioneer plaque and a 1948 study by information theorist Claude Shannon. They compressed the files into a master file, and then split the data into short strings of binary code made up of ones and zeros. Using an erasure-correcting algorithm called fountain codes, they randomly packaged the strings into so-called droplets, and mapped the ones and zeros in each droplet to the four nucleotide bases in DNA: A, G, C and T. The algorithm deleted letter combinations known to create errors, and added a barcode to each droplet to help reassemble the files later. In all, they generated a digital list of 72,000 DNA strands, each 200 bases long, and sent it in a text file to a San Francisco DNA-synthesis startup, Twist Bioscience, that specializes in turning digital data into biological data. Two weeks later, they received a vial holding a speck of DNA molecules. To retrieve their files, they used modern sequencing technology to read the DNA strands, followed by software to translate the genetic code back into binary. They recovered their files with zero errors, the study reports. (In this short demo, Erlich opens his archived operating system on a virtual machine and plays a game of Minesweeper to celebrate.) They also demonstrated that a virtually unlimited number of copies of the files could be created with their coding technique by multiplying their DNA sample through polymerase chain reaction (PCR), and that those copies, and even copies of their copies, and so on, could be recovered error-free. Finally, the researchers show that their coding strategy packs 215 petabytes of data on a single gram of DNA--100 times more than methods published by pioneering researchers George Church at Harvard, and Nick Goldman and Ewan Birney at the European Bioinformatics Institute. "We believe this is the highest-density data-storage device ever created," said Erlich. The capacity of DNA data-storage is theoretically limited to two binary digits for each nucleotide, but the biological constraints of DNA itself and the need to include redundant information to reassemble and read the fragments later reduces its capacity to 1.8 binary digits per nucleotide base. The team's insight was to apply fountain codes, a technique Erlich remembered from graduate school, to make the reading and writing process more efficient. With their DNA Fountain technique, Erlich and Zielinski pack an average of 1.6 bits into each base nucleotide. That's at least 60 percent more data than previously published methods, and close to the 1.8-bit limit. Cost still remains a barrier. The researchers spent $7,000 to synthesize the DNA they used to archive their 2 megabytes of data, and another $2,000 to read it. Though the price of DNA sequencing has fallen exponentially, there may not be the same demand for DNA synthesis, says Sri Kosuri, a biochemistry professor at UCLA who was not involved in the study. "Investors may not be willing to risk tons of money to bring costs down," he said. But the price of DNA synthesis can be vastly reduced if lower-quality molecules are produced, and coding strategies like DNA Fountain are used to fix molecular errors, says Erlich. "We can do more of the heavy lifting on the computer to take the burden off time-intensive molecular coding," he said. The Data Science Institute at Columbia University is training the next generation of data scientists and developing innovative technology to serve society. http://datascience. Columbia Engineering is one of the top engineering schools in the U.S. and one of the oldest in the nation. Based in New York City, the School offers programs to both undergraduate and graduate students who undertake a course of study leading to the bachelor's, master's, or doctoral degree in engineering and applied science. Columbia Engineering's nine departments offer 16 majors and more than 30 minors in engineering and the liberal arts, including an interdisciplinary minor in entrepreneurship with Columbia Business School. With facilities specifically designed and equipped to meet the laboratory and research needs of faculty and students, Columbia Engineering is home to a broad array of basic and advanced research installations, from the Columbia Nano Initiative and Data Science Institute to the Columbia Genome Center. These interdisciplinary centers in science and engineering, big data, nanoscience, and genomic research are leading the way in their respective fields while our engineers and scientists collaborate across the University to solve theoretical and practical problems in many other significant areas. The New York Genome Center is an independent, nonprofit academic research organization at the forefront of transforming biomedical research and clinical care with the mission of saving lives. A collaboration of renowned academic, medical and industry leaders across the globe, the New York Genome Center's goal is to translate genomic research into development of new treatments, therapies and therapeutics against human disease. Its member organizations and partners are united in this unprecedented collaboration of technology, science and medicine, designed to harness the power of innovation and discoveries to advance genomic services.
News Article | October 28, 2016
It's difficult to make predictions, especially about the future, and even more so when they involve the reactions of living cells -- huge numbers of genes, proteins and enzymes, embedded in complex pathways and feedback loops. Yet researchers at the University of California, Davis, Genome Center and Department of Computer Science are attempting just that, building a computer model that predicts the behavior of a single cell of the bacterium Escherichia coli. The results of their work were published Oct. 7 in the journal Nature Communications. The new simulation is the largest of its kind yet, said Ilias Tagkopoulos, professor of computer science at UC Davis, who led the team. "The number of layers, and the amount of data involved are unprecedented," he said. The dataset on which the model is based includes, for example, over 4,389 profiles of the expression of different genes and proteins across 649 different conditions. Both the dataset, named "Ecomics" and the integrated model, MOMA (Multi-Omics Model and Analytics) are available to other researchers to use and test. The model could be useful to researchers as a fast and inexpensive way to predict how an organism might behave in a specific experiment, Tagkopoulos said. Although no prediction can be as accurate as actually performing the experiment, this would help scientists design their hypotheses and experiments. Applications range from finding the best growth conditions in biotechnology to identifying key pathways for antibiotic and stress resistance. Collecting and downloading the data took a week, but processing the data into a single dataset took two years of the three-year project, Tagkopoulos said. The team built models for four layers, starting with gene expression and working up to the activity at the whole-cell level. Then they integrated the layers together. They used techniques in machine learning to train the models to predict the behavior of each layer, and ultimately of the cell itself, under different conditions. The model was built on computer clusters at UC Davis, and on supercomputers available through a national network. The researchers received a National Science Foundation grant of computing time on "Blue Waters," one of the world's most powerful supercomputers, at the National Center for Supercomputer Applications. Although E. coli is a well-known organism, we are far from knowing everything about its biochemistry and metabolism, Tagkopoulos said. "We are exploring a vast space here," he said. "Our aim is to create a crystal ball for the bacteria, which can help us decide what is the next experiment we should do to explore this space better." With collaborators at Mars Inc. Tagkopoulos hopes to begin building similar databases and models for bacteria involved in foodborne illness, such as Salmonella enterica and Bacillus subtilis. He expects other researchers to draw on the Ecomics database, and hopes to make the MOMA model interface more accessible for biologists to use. "We're living in an amazing era at the intersection of computer science, engineering and biology," he said. "It's a very interesting time." Co-authors on the paper are Minseung Kim at the UC Davis Department of Computer Science and Genome Center, and Navneet Rai and Violeta Zorraquino, UC Davis Genome Center. The work was supported by the U.S. Army Research Office and the National Science Foundation.
News Article | March 30, 2016
No statistical methods were used to predetermine sample size. The investigators were not blinded to allocation during experiments and outcome assessment. Platynereis dumerilii embryos were collected in EMBL Heidelberg, Germany. In each of several containers, a gravid male and a female were mixed in a small container containing North Sea water. The classical breeding dance was observed after several minutes and the females and males released oocytes and sperm. Fertilized eggs were incubated at 18 °C and embryos were collected every hour after fertilization for a period of 5 days. Individual embryos were collected on the cap of a 1.5 ml Eppendorf tube using a micro mouth pipette. Excessive water was removed and the sample was flash-frozen in liquid nitrogen. Schmidtea polychroa embryos were collected at the Max Planck Institute CBG, Dresden, Germany. A population of S. polychroa was maintained in the lab at 20 °C as previously described31. Egg capsules were regularly collected over 15 days of development just after deposition and kept in Petri dishes at 20 °C. To release embryos for isolation, capsules were carefully opened using two fine forceps. After assessment of stage of development according to the Martin–Duran system32 excess water was removed and embryos were flash-frozen in Eppendorf tubes in liquid nitrogen. Hypsibius dujardini starting cultures were provided by Bob Goldstein (University of North Carolina at Chapel Hill) and embryos were collected as previously described33 at the Technion, Israel. Small cultures of tardigrades were kept in 60mm glass Petri dishes in commercial bottled spring water until gravid animals were visible. Tardigrades lay 2–5 eggs during molting, with the embryos deposited in their shed exoskeleton, the exuvia. Soon after the adult crawled out of the exuvia, it was cut open using a scalpel on a microscope cavity-slide to release embryos into the medium. Embryos were observed using a standard binocular and when reaching two-cell stage were deposited in a 10 μl drop mineral water on the cap of a 1.5 ml Eppendorf tube. Tubes were incubated for respective periods at 20 °C. For 4.5 days, once per hour past the two-cell stage, embryos were inspected for viability, excessive water was removed using a micro mouth pipette, and the tube was flash-frozen in liquid nitrogen. Drosophila melanogaster embryos were collected at the Technion using a previously published protocol34. Briefly, agar plates with apple juice smeared with freshly prepared yeast were used to make young adult flies lay a lot of eggs. Cages consisting of such plates were set up with at least 20 flies and left for roughly one day for the flies to acclimatize. Plates were replaced with fresh ones twice in one hour interval to ensure the use of only newly laid eggs. Drosophila embryos are covered with a non-transparent chorion which has to be removed before live imaging by dechorionation. Shortly after being laid, embryos were washed off from plates into a plastic sieve using tap water and a fine brush used to loosen the embryos. In the sieve, embryos were submerged in 50% bleach solution for two minutes. Embryos were washed and rinsed with cold water. Using a needle pick, 20 embryos were placed in a row on a strip of agar placed on a glass slide. n-Heptane glue was applied in a thin layer on a big glass cover slip. This coverslip was carefully put upside down on top of the embryos on the agar strip so that embryos adhere to the glue layer. Embryos were covered with PBS and kept in humid chambers at 25 °C. Embryos on the coverslip were observed under the light microscope and for each embryo, the time of cellularization was noted. To collect an embryo, a needle pick was used to carefully remove it from the slide and it was flash-frozen in an Eppendorf tube in liquid nitrogen. Strongylocentrotus purpuratus oocytes and sperm were kindly provided by Smadar Ben Tabou deLeon (Haifa University, Israel) and mixed and cultured at the Technion. Mixing occurred by 4 drops of sperm in 50 ml of eggs in sea water and incubated at 18 °C in Petri dishes. After fertilization, every 40 min (for a period of 72 hours), single embryos were deposited on the cap of a 1.5 ml Eppendorf tube. Excessive water was removed using a micro mouth pipette and the embryo was flash-frozen in liquid nitrogen. Danio rerio fertilization was performed in the lab of Karina Yaniv (Weizmann Institute, Israel). Four female and one male Danio rerio fish were mixed in a breeder tank. Fertilized eggs were collected into zebrafish embryo medium as previously described35. Fertilized eggs were sampled in a small volume of medium every 40 min from fertilization into the cap of a 1.5 ml Eppendorf tube. Excess water was removed using a micro mouth pipette and the embryo flash-frozen in liquid nitrogen. Nematostella vectensis egg masses and sperm were provided by Amos Schaffer (Gat Lab, Hebrew University of Jerusalem). Eggs and sperms were mixed and egg jelly was dissolved as previously described36 using 4% cysteine (pH 7.4–7.6) to make single embryos accessible for collection. The embryos were washed with cysteine six times using 30% of sea water. Fertilized embryos were observed under a light microscope and embryos reaching the 4-cell stage were deposited in a 10 μl drop of 30% salt water on the cap of a 1.5 ml Eppendorf tube. Tubes were closed and incubated for respective periods at 20 °C. At collection time, embryos were inspected for viability and excessive water was removed using a micro mouth pipette. The embryo was then flash-frozen in liquid nitrogen. After (and including) the four-cell stage, every 20 minutes (for a period of 48 hours when the embryos reached the late planula stage), single embryos were deposited on the cap of a 1.5 ml Eppendorf tube. Mnemiopsis leidyi embryos were collected in the Whitney Institute, University of Florida as previously described37. Stages ranged from the fertilized egg to 20 h. Three replicate time-courses each comprising 20 embryos were isolated. In one replicate, embryos were flash frozen and shipped on dry ice. In the other two RNA was prepared by a TRIzol extraction and shipped in 75% ethanol on dry ice. For Hypsibius dujardini, Schmidtea polychroa, and Platynereis dumerilii, RNA was isolated from a mixed population of embryos, larvae, and adults according to the TRIzol protocol (Invitrogen). This RNA was processed according to the Illumina TruSeq RNA-seq protocol by the Technion Genome Center and 100 bp paired-end sequencing was performed. To pre-process the reads, ‘Sickle’38 was used for quality trimming with a threshold of 31 and Illumina adaptors were removed using ‘Scythe’39. Sequencing error correction was next made using the AllpathLG toolkit40 and poly-A sequences were trimmed using trimest (Gary Williams, unpublished). The resulting libraries were cleaned of short and duplicate reads using the fastx toolkit (Assaf Gordon, unpublished). Hypsibius dujardini’s genome has been recently reported by two groups41, 42. The sea urchin transcriptome was downloaded from Echinobase, NCBI BioProject PRJNA81157. The longest Isoform per transcript was selected leaving ~21,000 peptides. For this organism the mapping was done more loosely with bowtie parameters set to “–mp 3,1 -N 1 -L 15” as the RNA-seq was done on a heterogenic population. For Nematostella we retrieved the T1 transcriptome from Stellabase43. Using Transdecoder (https://transdecoder.github.io) revealed that, of the ~115,000 transcripts, only ~53,000 encoded proteins. BLAST analysis of the encoded protein resulted in ~42,000 unique proteins and the longest transcript was selected for each protein. Total RNA was extracted from single embryos using TRIzol as previously described7 including minor adjustments. After the addition of TRIzol to the embryos the mixture was frozen in liquid nitrogen, thawed at 37 °C and vortexed for 30 s. This procedure was repeated five times. Chloroform was then added and the sample further processed. The dried total RNA pellet was dissolved in RNase-free water before introduction into subsequent amplification and sequencing library preparation steps. Using the CEL-Seq protocol44, 1 μl of a single embryo total RNA sample with a maximum concentration of 50 ng μl−1, was mixed with 1 μl of the ERCC spike-in kit diluted according to the manufacturer’s protocol45. The libraries were sequenced using Illumina paired-end sequencing as previously reported in the CEL-Seq protocol44. For Read 1, used to determine the barcode, the first 15 bp were sequenced and for Read 2, used to determine the identity of the transcript, the first 35 bp were sequenced. The CEL-Seq pipeline is available at https://github.com/yanailab/CEL-Seq-pipeline. Transcript abundances were obtained from the sequencing data using custom scripts organized into a multistep paralleled computational pipeline. Briefly, after trimming and filtering, the paired-end reads were demultiplexed based on the first eight bases of the first read. For each sample, reads were mapped to a reference genome or transcriptome using bowtie2 version 2.2.3 (ref. 46) with default parameters and counted using htseq-count47 to generate read counts. Samples were filtered to include only samples with at least 500,000 reads and in additions ERCC spike-in information was also used to filter out samples with low correlation coefficients (<0.65) to the known concentration or with high (>0.3) spike-in to gene read count ratio. Read counts were then normalized by dividing by the total number of counted reads and multiplying by 106. Because CEL-Seq retains only the 3′ end of the transcript, this procedure yields the estimated gene expression levels in transcripts per million (tpm) without transcript length normalization. In this work, we compare the transcripts per million developmental profiles for different genes and across orthologues, and such comparisons are generally robust to overall RNA content changes. A de novo transcriptome was generated for S. polychroa, P. dumerilii and H. dujardinii. Since we had at our disposal CEL-Seq reads, in addition to the RNA-Seq reads, our strategy was to exploit the stranded and 3′-biased nature of CEL-Seq. The Trinity software48 was used to generate, for each of the three species, two de novo transcriptome assemblies: (1) single-end CEL-Seq reads were used to generate a 3′ biased stranded transcriptome, and (2) the CEL-Seq reads were combined with paired-end RNA-seq reads were used to generate a combined transcriptome. For the CEL-Seq 3′ assembly, we ran Trinity using the single-end mode with ‘SS_lib_type’ parameter set to ‘F’. For the combined assembly we ran Trinity using the paired-end mode with default parameters. The two resulting transcriptomes were then used to generate a single 3′ anchored stranded transcriptome. For each transcript (contig) in the first set, we identified the corresponding transcripts in the second set using BLAST49. Of those identified, we selected the transcript with the highest alignment score and used the strand information of the transcript in the first set to generate a stranded transcript (Extended Data Fig. 1). Genes with alternative 3′-ends may be represented as distinct genes in this set, in those rare cases when the CEL-Seq contigs do not overlap. The generated set of transcripts was further filtered to contain only transcripts with a predicted protein using the Trinotate pipeline that is a part of the Trinity software48. PFAM domains50 were then identified using HMMER51. GO annotations for each transcriptome were generated using Trinotate (http://trinotate.github.io/). Specifically, transcripts were searched against Uniprot sequences (comprising SwissProt and Trembl invertebrate, vertebrate, mammal, rodents and human data, clustered to 90% identity). GO and PFAM identifiers were then extracted from Uniprot accessions. OrthoMCL52 was used to delineate orthologous clusters from the ten proteomes of the ten species using the following parameters: “percentMatchCutoff” was set to 24, “evalueExponentCutoff” was set to −5, and the MCL parameters were “–abc -I 1.5”. In the case of multiple genes in an orthology cluster for a particular species, the one with the highest fold-change was selected as the representative. We found similar results if the representative is selected randomly among the inparalogues. Each time-course was initially ordered using BLIND—an automated method for determining the developmental order of transcriptomic samples13 (Extended Data Fig. 2a). These profiles were smoothed using a moving average calculation with span parameter set to 3. In order to compare profiles of equal lengths, for each species we reduced the time-course to twenty sliding windows using the following method. We defined the size of the window such that there is only overlap between every two consecutive windows. For each window, the average expression was calculated for each gene across the included embryos. For each time-course, dynamic genes were defined as those with minimum expression of 10 transcripts per million and at least a twofold change. Standardized expression was used in analysis where noted: to generate a standardized expression, the mean expression value was subtracted from each expression value and the results were divided by the standard deviation. To generate the phasegrams shown in Fig. 2 we first standardized the log profiles by subtracting the mean and dividing by the standard deviation. We next computed the first two principal components of this expression data; since the profiles were standardized, the genes form a circle. The genes are then sorted according to their angle from the origin in this space. A gene expression profile was mapped to a temporal phase (early, transition, or late) by computing the correlation with the three idealized profiles shown in Extended Data Fig. 5 and assigning it to the pattern exhibiting the highest correlation and thus best match. The transition period for each species was computed based upon the transcriptome similarities with the transcriptomes of the other species, shown in Fig. 3. The twenty transcriptomes were clustered using hierarchical clustering based upon the Euclidean distances among their profiles of correlations with the profiles of all other species. The two deepest clusters were then identified and the precise temporal window separating them was set as the mid-developmental transition period. A temporal phase was assigned to each orthologous group by annotating it to its most represented phase. The C. elegans Gene Ontology annotation was used on the C. elegans orthologues. Enrichment was computed using the hypergeometric distribution. In order to avoid retrieving enrichments due to the same set of genes we carried out serial enrichments as follows. The most enriched gene ontology group was noted, its genes removed from the set, and enrichment search was repeated to detect additional Gene Ontology terms. For the signalling pathways shown in Fig. 4b, the following gene ontology terms were used: ‘Wnt signalling pathway’, ‘Notch signalling pathway’, ‘hedgehog receptor activity’, ‘epidermal growth factor receptor signalling pathway’, ‘transforming growth factor beta receptor signalling pathway’, ‘MAPK cascade’, ‘G-protein coupled receptor activity’. For this analysis, we searched for enrichment up to three windows before and after the inferred transition, and kept the most significant P value for each pathway (hypergeometric distribution). For each of 5,745 PFAMs, we computed an enrichment profile throughout time, and for each species, as follows. For each of the twenty expression windows of the matrix of standardized log expression values of the dynamic genes, we marked genes with expression above 0.5 as expressed. We then calculated the fraction of the genes within this set that contain genes annotated to the PFAM domain. A temporal phase was annotated using supervised clustering using the same approach shown in Extended Data Fig. 5. For the transcription factor families shown in Fig. 4d the following PFAMs were used: ‘Homeobox domain’, ‘GATA zinc finger’, ‘Ligand-binding domain of nuclear hormone receptor’, ‘Helix–loop–helix DNA-binding domain’, ‘bZIP transcription factor’, ‘Zinc finger, C4 type (two domains)’, ‘Zinc finger, C2H2 type’, and ‘T-box’. For this analysis, we searched for enrichment up to three windows before and after the inferred transition, and kept the most significant P value for each TF family (hypergeometric distribution).
News Article | December 5, 2016
The ability to convert cells from one type to another holds great promise for engineering cells and tissues for therapeutic application, and the new Wisconsin study could help speed research and bring the technology to the clinic faster. The new approach, published this week (Dec. 5, 2016) in the Proceedings of the National Academy of Sciences (PNAS), uses a library of artificial transcription factors to switch on genes that convert cells from one type to another. Natural transcription factors are cellular molecules that bind to DNA to turn genes on and off. They help determine cell fate, meaning that if a cell is destined to be a skin cell, a heart cell or an eye cell, different transcription factors switch on specific sets of genes that program the cell to attain one state or another. Using artificial transcription factors made in the lab, researchers are trying to find which ones best mimic these natural changes in cell fate. "Our interest in changing cell fate comes from understanding how cells selectively use the information in our genomes to make specific cell types and also from the many therapeutic benefits such knowledge can offer," says Asuka Eguchi, the study's lead author and a member of Professor Aseem Ansari's lab in the UW-Madison Department of Biochemistry. "For example, if a patient needs a certain cell type, the idea is we can reprogram their own cells to what they need, rather than relying on donor cells. This allows us to study patient-specific cells and potentially avoids issues with immune response where a patient's body could reject the cells." Conventional methods for finding the correct factors to change cell fate require scientists to perform a trial-and-error approach. They need prior knowledge about which combination of the thousands of natural factors would possibly work within a tightly choreographed timeframe to program cell fate. It is a slow, laborious, failure-prone process, the researchers say. The new method utilizes "libraries" of millions of artificial transcription factors that were designed to bypass natural controls and switch on genes that might be activated in a given cell type. In addition, the factors contain an attachment that lets them bind and work in concert to affect genes, a step not traditionally taken. By exposing the library of factors to cells, they can see if cell fate changed in any of them. If so, they can revisit those cells to see which factors were responsible. For their experiments, the Wisconsin group started with mice fibroblasts, a cell in connective tissue, and looked for them to be reprogrammed into what are called induced pluripotent stem cells. Given proper cues, these types of stem cells can become any type of cell in an animal's body, including humans. By reprogramming, the researchers mean that the artificial factors would trigger all of the right genes to cause the cell to shift from one type to another. "Imagine you have millions of keys and only a unique key or combination of keys can turn a motor on," says Ansari, who is also affiliated with UW-Madison's Genome Center of Wisconsin. "We test all those keys in parallel and when we see the motor fire up, we go back to see exactly which key switched it on." In the process of testing their tool, the researchers discovered three combinations of the artificial factors that reprogrammed a fibroblast into a stem cell. The factors played a role similar to that of a natural transcription factor important in a process, called Oct4. "In this unbiased approach, we can try to basically cast a wide net on the whole genome and let the cell tell us if there are important genes perturbed," Ansari says. "It's a way to induce cell fate conversions without having to know what genes might be important because we are able to test so many by using an unbiased library of molecules that can search nearly every corner of the genome." The reprogramming of fibroblasts into stem cells has been well studied. The researchers put their approach to the test in this context because it places a high-bar and requires significant changes to the cell. With this proof of concept, the Wisconsin scientists hope other researchers use their method to discover new genes that can drive more difficult conversions of cell fate. "Generating these pluripotent stem cells also helps us avoid having to make embryonic stem cells, which can be controversial," says Eguchi, who is a recent graduate of the UW-Madison Cellular and Molecular Biology Training Program. "We can also start better investigating direct conversions, which are conversions from one cell type to another without the need to go to the pluripotent stage first because that can cause problems in some contexts. This tool opens up the doors to research these areas more effectively." Explore further: Removing cellular bookmarks smooths the path to stem cells More information: Reprogramming cell fate with a genome-scale library of artificial transcription factors, www.pnas.org/cgi/doi/10.1073/pnas.1611142114