Sure, Li. Thanks for the questions. So on the update for 007, as we've said, we've dosed at this point more than 100 patients in the ongoing 007 study. Just to remind everyone, we need to dose a minimum of 72 patients across the four different genetically driven cohorts at each of 200, 400, and 600 milligrams. So that's, four genotypes times three doses times at least six patients per dose. The reason we've enrolled more than the 72 is, we think, due to investigator and patient enthusiasm to get access to ziftomenib, as well as the fact that we leave these cohorts open for enrollment while we're moving to the next dose level. So patients that are in screening are eligible to go into that cohort if the next cohort hasn't yet opened. So we've actually seen, I believe, over-enrollment in nearly every cohort at every dose. Li, what we're encouraged by is the fact that the safety and tolerability continue to be very consistent with what you saw in our January update. I think enrollment has been robust, and we would expect once we get into these expansion cohorts, enrollment should continue to be robust because we are, in the case of 7+3, removing the restrictions on adverse risk. I'll just remind everyone the escalation is in the adverse risk population. And in the case of Ven/Aza, we will move from the relapsed/refractory setting to expand in both the frontline Ven/Aza as well as to expand in the relapsed/refractory setting. So by the time we get to the end of the year, Li, we'll - that's why I say, I don't think we know exactly today what the update will be completely, but it's going to be a meaningful update with a lot of patients, a lot of data, good durability across the cohorts. Obviously, the 200-milligram cohorts having been on the longest, we're looking forward to sharing that update when the time comes. In terms of GIST, you asked about the gating steps for the trial. At this point, the study is in study startup. So, we'll do this with relatively few U.S. sites initially. Now that we have an open IND, we can move as quickly as possible through the site initiation and contracting phase. Things have changed quite a bit from when I started in this industry. You used to be able to do this in a couple of months. It's a bit longer now to get up and running. We do think, first half next year is a reasonable guide for first patient in. We use - we worked with multiple leading KOLs in GIST, both to evaluate the data and to help design the study. You asked, what does success look like? That really came from those KOLs. So we're going to deliberately be in the population that is progressing or has just failed imatinib. So what you're looking to do is to reverse that. Again, you'll see that the nonclinical data, it goes quite a bit beyond what Dr. Armstrong and his colleagues presented in their paper. But if you can, either - if you can drive responses, right, durable responses, that would be the gold standard. That's really what we're looking for. And we are going to do some amount of dose optimization to ensure that we meet the requirements of the FDA's Project Optimus initiative, right? That continues as you move through these combos. So we'll talk more about, Li, kind of - once we show you the nonclinical data, we can talk a little bit more about what to expect. Obviously, if we get a good signal, our goal would be to go straight into the frontline. Because this combination is so powerful and so orthogonal to how people have treated GIST, we think it really could be transformational for those patients and really help, drive durable responses. But let's take it one step at a time.