Evolution of Gene Expression in Drosophila embryogenesis
–Hourglass model
Alex T. Kalinka, … Casey Bergman
: use microarray to interrogate the expression in multiple species, use mixed anova to look at time class * species interaction. What they found is an hourglass shape, meaning that genes expressed early or late are more likely to diverge in expression than the middle, which is around stage 5. The hypothesis is that things in the middle there is some bottleneck effect.
Evolutionary relics of TFBS turnover
Albert Erives and Justin Crocker
: in the NEEs, comparing mel to ana/wil which have uncompacted genomes
- they gradually loosen the criteria for defining a dorsal site, as a result they defined a “cis-spectra”. the idea is that some of them which are most degenerate may not be considered a binding site by most people studying enhancer. But for their hypothesis, these elements represent relics of evolutionary process
- the Da site is even considered a “mirage”: the hypothesis is that dorsal site has moved so many times so as to leave its traces everywhere. Combined with some motif overlap between the two sites, it becomes very likely to evolve a Su(H) site on the base of a dorsal site relic. For young NEE that hasn’t had a background of dorsal relics, the Su(H) doesn’t look like that.
- There hasn’t been functional assays to show that the rest of the sites are non-functional, but the hypothesis is that they just represent historical relic
- one thing to notice about NEE: in this class of enhancer, dorsal sites work in conjunction with the E(CA)T, and possibly only by collaborating with E(CA)T will the dorsal site be functional. Thus the enhancer doesn’t care if it has “additional binding site relics”… Are there evolution against spurious sites, so that as binding sites move, their traces are actually erased?
Eisen, ChIP-seq
: the sequenced tags are from both ends, so they need to be extended to match the actual piece; the MACS output hasn’t been normalized across genomes. But for my purpose I could use the #tag as a proxy for maximal binding strength (some sort of fragment density)
Lectrin-42a, molecular evolution in drosophila in response to parasitic wasp
Erin S. Keebaugh, Todd A. Schlenke
: These wasps lay eggs in drosophila embryos, the later has an immune response which will form a wasp bag to exclude the wasp egg. But not always successful. They found by comparing several populations of simulans that a new gene lectern 42a has arisen, they have evidence from selective sweep, MK test and dN/dS to show that it has been driven by positive selection.
Dosage compensation and demasculinization of X chromosome
Doris Bachtrog
: raise a 3rd possible mechanism. in drosophila dosage compensation complex binds to certain locus (sequence specific) on the X chromosome in males. She had two alternative hypothesis: either the binding allows the gene to have normal expression level in males so that male-biased gene are more likely to arise there; or the binding changes the local chromatin structure such that male biased gene expression patterns are harder to evolve. She showed that the later hypothesis is supported by data. Not only are the male biased genes increase in number as going away from high affinity sites for the complex, the level of male biased expression also go up.
PIWI-RNA
Lu, Jian and Andy Clark
: Jian showed that PIWI RNA could effectively reduce TE expression. He classified two classes of TE insertion: those that insert into PIWI locus, which then become inactivated because it will generate a PIWI RNA, and those that not, which will be PIWI RNA targets. He showed using simulation that PIWI RNA will effectively make the target TE less deleterious and thus natural selection won’t be as harsh on them. So on one hand PIWI RNA “detoxify” TE, on the other hand, these TE can accumulate more readily and consequently the fly will become dependent / “addicted” to the PIWI RNA.
TF binding affinity evolution
Colin W. Brown, Mike Eisen
: their approach is 1) look for sequence alignment of TF to look for candidates; 2) for candidate TF they did chip-seq in the species where the TF differ in sequence to look for binding differences. They found one candidate, her (for hermaphrodite), which is a C2H2 TF, in pse it has 2 more zinc fingers than mel. using chip-seq and motif analysis, they show that mel has 302 targets and pse has 74 targets, with only 4 overlaps. The motif analysis showed that pse her has more specificity in the middle where mel her is not very selective. The functional consequences of this change is awaiting further investigation.
Note that another student in Eisen lab is working on similar projects.
Shadow enhancer confer robustness
Alistair N. Boettiger, Mike Levine
: 1) through chip-chip experiments they found lots of additional peaks near the “expected” peaks (in known enhancers and overlapping known sites). cloning the sequences containing these additional peaks found quite a few potential enhancers that have nearly identical expression patterns as the primary one.
2) he showed that a construct containing both the primary and the shadow enhancer lead to ~80% of the cells expressing, while only the primary or the shadow lead to ~40% expressing.
3) do both enhancers function at the same time in one cell or is it more like a backup mechanism? do they arise and are maintained by selection? two is always better than one?
Trans-level redundancy for robust expression
Robert J. Johnston, … Claude Desplan
: fly eye,
Runt-dependent regulation of enhancer-promoter interactions
J. Peter Gergen, Lisa Prazak
: implications for enhancer people–cannot remember. but the work has to do with two cis modules that have some dependency on each other. One element can loop and work as a silencer of the other…
RNAi
TRiP–Harvard Medical School
VALIUM vectors
Ni, J-Q
Drosophila Piwi functions in Hsp83-mediated capacitance for morphological evolution.
Vamsi Gangaraju, Haifan Lin
: Very interesting work. The problem stems from the observation that knock down of Hsp90 reveals cryptic variation and the revealed variation become further inheritable even in the absence of Hsp90 mutation, suggesting the later functions by leaving some trace on the chromosome. A previous nature paper in 2010 suggests that the phenomenon is caused by the loss of piwi which normally has TE inhibition functions. They show that knock down of the piwi RNA pathway can phenocopy the Hsp- experiments. Further they ask the relationship between Hsp and piwi RNA–do they function in parallel pathway or is one upstream of another in the same pathway? They were able to rule out either piwi upstream of Hsp or parallel hypothesis, which leave them to Hsp functioning through piwi hypothesis. They also argue against the aforementioned explanation involving piwi inhibiting TE, as they didn’t observe an elevated TE transcription level (or actual transposition rate?) in piwi knockout strain.
in toto imaging of Drosophila embryo
Willy Supatto, Thai Truong
: Similar techniques with last year’s breakthrough in Science, on imaging whole zebrafish embryo development. This group improved the one photon light sheet microscopy by switching to 2 photon–use GFP as an example, instead of using one photon of blue light to excite the fluorescence, use 2 photons in far red spectrum which sums to the required excitation energy. Because far red spectrum are much less scattered by the tissues (and thus less absorbed), it is both more penetrant and also less phototoxic. Using this technique they could image the embryo development both to high resolution and for long time (24hr).
// evidence of changes in the cabin pressure! My water bottle which has an air-tight seal, containing no liquid but air before boarding, suddenly burst open (with a cracking sound). The lid ejected out. One explanation is that the cabin pressure just changed to the amount that, although we didn’t notice, is strong enough to create a pressure difference between inside and outside the bottle. The inside is bigger and so ejects the lid.
Increased cell bond tension governs the cell sorting at the Drosophila anterior-posterior compartment boundary.
Daiki Umetsu, Christian Dahmann
: during development, cells with different fates may form a sharp boundary, like in drosophila’s anterior-posterior compartments. Simulation show that if cell replication is random, this boundary will be “blurred” over time, which is not consistent with the observation. So the question is how are cell boundaries maintained during cell proliferation. The model they come up with is through between cell bonds. They reasoned that if cell-cell bond tension is higher at the boundary, this will in general bias the direction of boundary cell replication towards its own side (it’s harder to cross the tension line, which I haven’t fully understood yet). They used laser to cut the cell boundary in order to measure the bond strength (by measuring the vertex displacement after cutting). They show that there is a 2.5 fold increase in cell-bond at the boundary. In simulation as they incorporate this 2.5 fold difference, the boundary is significantly more straight than without. However the simulation is still not as good as what is observed in nature, suggesting other mechanisms operating.
Low affinity Ci-sites
Scott Barolo
: there are two types of Hedgehog target enhancers–those that respond throughout the tissues and those that respond in a tissue specific manner. The former features Ci sites that are perfect to the consensus. The later, however, consistently contain sub-optimal sites (2-4 bp away from the 9 bp motif). Not only is this the case in mel, the phenomenon is conserved across species. Functional experiments show that optimizing these “bad” Ci sites to optimal sites often cause counterintuitive results, which they realized may have to do with Ci protein’s dual function–it works as an activator at one concentration and then dimerize to become a repressor at another. This creates a non-monotonic response to Ci sites.
Misha’s MSE robustness experiments
: the question is when MSE is put into the 16 kb construct replacing the bigger S2E region, i.e. deleting the flanking 300 bp without putting back some sort of spacers, it may have brought MSE in close contact with nearby enhancers such as S3E, inducing unexpected results. There are fitness reduction in this case, which differs between males and females.
Adaptive substitutions in Drosophila melanogaster
Bryan Kolaczkovsky, Andrew Kern
: the first study I heard of using DPGP dataset. Used a modified compound test to detect selection in coding regions. They found both standing genetic variations contributing to the adaptation and also new mutations.
Evolutionary mirage
Lusk and Eisen
: One of my questions is that overlapping site should be functionally distinct from two non-overlapping sites due to competition for binding. Eisen pointed out (or rather suggested) that occupancy is often low such that competition may not be often occurring. The evidence he used is that in their modeling gene expression attempts, incorporating or not competition doesn’t influence the result to a large extent.
Susan’s poster on dosage compensation for early zygotically expressed genes
: the motivation is that Gt, a zygotic gap gene is on X chromosome. And development process is expected to be sensitive to the dosage of Gt. Thus it becomes a question of when dosage compensation mechanism is installed in place during early embryogenesis. She approached this question by RNA-seq on single embryo (which allowed her to precisely stage the embryo by taking a nuclei picture). She found that GT is properly balanced. However as she looked at the whole genome pattern, surprisingly, it turned out that Gt is an except
on rather than the rule.
Poster–Quantitative imaging of the dorsal nuclei gradient reveals limitations to threshold-dependent patterning in Drosophila
Gregory T. Reeves, Angelike Stathopoulos
: as a morphogen gradient which controls many genes’ expression, dorsal expression pattern does not reach steady state during its functional stage, suggesting that “thresholding” events happened while dorsal gradient is still dynamically changing and unstable. This creates a problem for previous models which rely on precise morphogen gradient at their steady state to achieve precise patterning. — this leads me to wonder: if TF gradients themselves during embryogenesis is non-steady, then does having precise amount of binding sites matter? Manu’s thesis land on this problem: threshold is a convenient but incorrect concept. robust patterning maybe achieved through gap gene cross-regulation! read Manu’s work.
