When Biology Meets Computer Science
I actually never planned to go into computer science for a living. If you had told me growing up as a little girl that was going to be my career path, I would have really thought you'd lost your mind. Um, I was interested in biology because of the impact on human health and I wanted to figure out how I could help cure different diseases. And I was really excited about that I only got into computer science out of necessity. It was during my PHD that I started on a project that required capturing a lot of images on the microscope. Basically me sitting at the microscope for four hours a day, incredibly tedious, incredibly boring. And I thought well this this microscope comes with some software that allows you to program it and I ended up creating this monster of a blob of code that allowed it to capture images and analyze them that really got me hooked. Seeing the power of being able to accelerate the biology was really enticing to me. I run a research lab here at the broad institute of Harvard and MIT and what's
really unusual about it is that we bring together people at the interface of two disciplines. So normally there's biologists and there's computer scientists and we are right in between those two. It's a really cool place to be because it's uncharted territory and the synergy of everyone coming together with their different ideas from different disciplines really helps to accelerate our research. So I got into this field because I wanted to develop new medicines. And it's an increasingly important problem as it's getting more and more expensive to find new medicines. And on top of that humanity is finding more and more diseases that we need to cure. The standard way of identifying drugs is that you just study a disease for a couple of decades. Try to figure out the
mechanisms there's so much that's unknown and it's just incredibly painstaking. There's still a huge bottleneck in finding medicines for each disease. Instead of trying to study diseases in actual humans or even in model systems like a mouse or a rat what if we could study the disorders in cells growing in a dish. They're super cheap, very easy to grow. And they don't mind if you add thousands of chemicals to them to try to figure out if one of them might make a good drug. So the approach that my lab is taking involves images. So we use microscopy, we use image analysis to see whether we can identify any unusual structures in the cells of humans that have particular disorders. We can treat those cells with different drugs and try to find are there any drugs that make the unhealthy looking cells look healthy again. One example of this basic approach is what we've done with patients who have various mental illnesses. And so what we thought is well, let's collaborate with someone here in the Boston area at
McLean hospital, gathering groups of patients that have very specific types of symptoms. And they've taken cells from those patients skin cells and grown them in a dish. And then we've looked at their structure. And we were really surprised to find that we could identify whether a patient has various illnesses that cause psychosis just by the structure of their cells, really a shocking finding! We're really excited to see that the mitochondrion of the cell, the powerhouses of the cell are not as not distributed as evenly as they usually are when patients have these different forms of psychosis. And this was an incredibly exciting result
because it means we can now take cells that have this disordered looking mitochondrion and test drugs one by one using just cells in a dish. This field of image-based profiling to see that field really blossom and develop over the past decade has been
incredibly satisfying. So when a clinical trial comes out and our software has been used to identify a drug for it, there's nothing like that feeling.
当生物学遇上计算机科学
事实上，我从来没有打算以计算机科学为生。如果你告诉我，作为一个小女孩长大后，这将成为我的职业道路，我真的以为你疯了。嗯，我对生物学感兴趣是因为它对人类健康的影响，我想弄清楚如何帮助治愈不同的疾病。我真的很兴奋，因为我只是出于需要才进入计算机科学。在攻读博士学位期间，我开始了一个需要在显微镜上捕获大量图像的项目。基本上我每天在显微镜前坐四个小时，非常乏味，非常无聊。我认为这台显微镜附带了一些软件，可以让你对其进行编程，最后我创建了这个怪物般的代码块，让它能够捕获图像并分析它们，这真的让我着迷。看到能够加速生物学的力量对我来说真的很有吸引力。我在哈佛大学和麻省理工学院的广泛研究所经营一个研究实验室，真正不寻常的是，我们将两个学科交叉的人聚集在一起。所以通常有生物学家和计算机科学家，而我们就介于这两者之间。这是一个非常酷的地方，因为它是一个未知的领域，每个人带着来自不同学科的不同想法聚集在一起的协同作用确实有助于加速我们的研究。所以我进入这个领域是因为我想开发新药。随着寻找新药的成本越来越高，这个问题变得越来越重要。除此之外，人类还发现越来越多的疾病需要治愈。识别药物的标准方法是研究一种疾病几十年。尝试找出
机制中有太多未知的东西，而且非常费力。寻找针对每种疾病的药物仍然存在巨大的瓶颈。如果我们可以研究培养皿中生长的细胞的疾病，而不是试图研究实际人类或什至小鼠或大鼠等模型系统中的疾病，那会怎样呢？它们超级便宜，非常容易种植。他们不介意你向其中添加数千种化学物质，试图找出其中一种是否可以制成一种好药。所以我的实验室正在采取的方法涉及图像。因此，我们使用显微镜，使用图像分析来查看是否可以识别患有特定疾病的人类细胞中的任何异常结构。我们可以用不同的药物治疗这些细胞，并尝试寻找是否有任何药物可以使看起来不健康的细胞再次看起来健康。这种基本方法的一个例子是我们对患有各种精神疾病的患者所做的事情。所以我们认为很好，让我们与波士顿地区的某人合作.麦克莱恩医院聚集了具有非常特定症状的患者群体。他们从这些患者的皮肤细胞中取出细胞并将其培养在培养皿中。然后我们研究了它们的结构。我们非常惊讶地发现，我们可以通过患者的细胞结构来识别患者是否患有各种导致精神病的疾病，这真是一个令人震惊的发现！我们非常兴奋地看到，当患者患有这些不同形式的精神病时，细胞的线粒体（细胞的发电站）的分布并不像通常那样均匀。这是一个令人难以置信的令人兴奋的结果
因为这意味着我们现在可以取出具有这种看起来无序线粒体的细胞，并仅使用培养皿中的细胞来逐一测试药物。基于图像的分析领域，以了解该领域在过去十年中真正的蓬勃发展和发展令人难以置信的满足。因此，当临床试验出来并且我们的软件被用来识别药物时，没有什么比这种感觉更好的了。
用户评论