http://www.simplyhired.com/a/jobtrends/trend/q-bioinformatics Bioinformatics Job Trends This graph displays the percentage of jobs with your search terms anywhere in the job listing. Since April 2009, the following has occurred: • Bioinformatics jobs decreased 15%
v*g
32 楼
Bioinformatics Market to Grow at 26% CAGR by 2013 http://www.justprnews.com/3222/bioinformatics-market-to-grow-at-26-cagr-by-2013/ Bioinformatics offers an indispensable technology for function assignment and is widely used for gene annotation. Initial efforts in bioinformatics were focused on the analysis of DNA sequence data. However, presently, the scope and objectives of bioinformatics research and development have broadened owing to the high generation of data from various sources and for different cellular processes, continuously evolving analytical technologies, and increasing computational capability. The global bioinformatics industry has been witnessing double-digit growth rate for the past decade due to ongoing research. As per our new research report “Global Bioinformatics Market Outlook” the market for bioinformatics will surge at a CAGR of nearly 26% during 2011-2013.
d*r
33 楼
still not bad... in 2009, job listing in many areas decreased much more than 15%...
【在 v*******g 的大作中提到】 : http://www.simplyhired.com/a/jobtrends/trend/q-bioinformatics : Bioinformatics Job Trends : This graph displays the percentage of jobs with your search terms anywhere : in the job listing. Since April 2009, the following has occurred: : • Bioinformatics jobs decreased 15%
g*d
34 楼
Bioinformatics jobs decreased 15% !
anywhere
【在 v*******g 的大作中提到】 : http://www.simplyhired.com/a/jobtrends/trend/q-bioinformatics : Bioinformatics Job Trends : This graph displays the percentage of jobs with your search terms anywhere : in the job listing. Since April 2009, the following has occurred: : • Bioinformatics jobs decreased 15%
v*g
35 楼
Y-axis is proportion . This means bad.
【在 d*****r 的大作中提到】 : still not bad... : in 2009, job listing in many areas decreased much more than 15%...
hehe, no offense bioinfo faculty位置大膨胀 only exists in your dream NGS did create bunch of data analyst positions, but not many faculty positions. It is much easier for the existing bioinfo faculty members to jump into the NGS field and get funded... unless your school does not have any good bioinformaticians. Finally, there is a real demand for people who know the state-of-the-art high performance computing in the NGS field, but it is a pure CS stuff.
【在 w******y 的大作中提到】 : hehe, no offense : bioinfo faculty位置大膨胀 only exists in your dream : NGS did create bunch of data analyst positions, but not many faculty : positions. : It is much easier for the existing bioinfo faculty members to : jump into the NGS field and get funded... unless your school does not have : any good bioinformaticians. : Finally, there is a real demand for people who know : the state-of-the-art high performance computing in the NGS field, : but it is a pure CS stuff.
【在 w******y 的大作中提到】 : hehe, no offense : bioinfo faculty位置大膨胀 only exists in your dream : NGS did create bunch of data analyst positions, but not many faculty : positions. : It is much easier for the existing bioinfo faculty members to : jump into the NGS field and get funded... unless your school does not have : any good bioinformaticians. : Finally, there is a real demand for people who know : the state-of-the-art high performance computing in the NGS field, : but it is a pure CS stuff.
这里在讨论找工作的问题,不是有没有用的问题。 biostat的人都知道,其实统计在建模里面更撤但,很多东西只是数学上make sense, 里实际还差十万八千里,但这并不妨碍人家好找工作。 在美国,很多事情都是这样。看上去很撤的东西,只要做的人多了,就能赚到钱,没有 人关心make sense or not,象牙医就是。很多不用做的东西,你去看牙医,他们都推 荐你做,很多美国人也听他们的。律师也是。这就是这个国家的文化。金融招一大堆学 数学的,弄出无比复杂的公式,就是呼游你去买他们的产品,他们赚到钱就行了。
【在 d******y 的大作中提到】 : bioinfo is a joke for modeling in my opinion.
b*1
47 楼
Here is a quote on 2010 survey of scientist salaries. While the salaries of scientists drops in almost all fields, the salaries of bioinformatics scientists increased 5% from 2009 to 2010. This year’s Salary Survey saw drops in salaries across the board. Almost every specialty suffered a setback, some with dips as large as $20,000 ( ecology) and $28,000 (virology). A few select fields bucked the trend, however, posting salary increases this year: bioinformatics, biophysics, biotechnology, and neuroscience. It isn’t always easy to determine why these specialties saw salary raises while others saw cuts, but researchers in each field can speculate. “There have been increasing requirements in NIH RFAs for informatics components in large projects,” says Mark Musen, head of the Stanford Center for Biomedical Informatics Research at Stanford University. And the continued surge in high-throughput experiments has departments around the country increasing their demand for employees who can manage and interpret the data, which may be adding to the jump in bioinformatics salaries, adds Musen. “I’ve noticed this year that start-up packages for new faculty members in biomedical informatics have been enormously generous because the competition is so intense,” he says. “It doesn’t surprise me that [The Scientist’s] salary data reflect this situation.” Read more: Life Sciences Salary Survey 2010 - The Scientist - Magazine of the Life Sciences http://www.the-scientist.com/article/display/57788/#ixzz17Yn39NAo
b*1
48 楼
With the advent of high density chip technology, most biological experiments have been or will be performed in high-throughput. Tons of data are generated every day across the world, which are required to be deposited into some central repositories by most journals. For 70% human diseases, anyone can retrieve thousands of data across a variety of measurements from public repositories for free. There is a huge gap in between the data and application in clinics. No matter you are a bioinformatician or not, as long as you can make discovery from these data, validate them in new clinical samples, and translate them into clinics, there is huge market for you. I think that bioinformatics is transforming the biology and clinics. In the future, maybe all biologists gains bioinformatics skills and use it to make discoveries. Or bioinformatics scientists take over a large stake of biology . The name of bioinformatics might even disappear, but as long as you understand biological/clinic problems and know how to use informatics to find the data, solve the problem, you have a bright future.
b*1
49 楼
Bioinformatics was born from high-throughput data, and is evolving very fast. Try to ask 10 people on what is bioinformatics, and you might get 10 different answers. When more and more biology/clinics fields are accepting high-throughput technology and centralized repositories, bioinformatics is evolving with the market to occupy more and more fields. Bioinformatics scientists are from very diversified background, but one common feature we all share is that we all have interdisciplinary knowledge and are willing to take risk and efforts to enter novel and exciting fields.
d*r
50 楼
I think the bottle-neck is not bioinfo itself, or statistical methods/algorithm, or even math if we put aside the philosophical question about whether biology can be understood by math. The bottle-neck is high-throughput methods to acquire the data, including large-scale perturbing/imaging tools on molecular/cellular/animal level. This requires huge progress on instrumentation/nanotechnology/microfluidics/automation/material science/surface chemistry, etc. That's why Feng Zhang was so popular on the job market. It's also hard because this kind of tedious and non- hypothesis-driven work will be too hard to get funding. Only in some central-powered system like China can this be possible.
experiments deposited into anyone public as long clinical you. I the
【在 b*******1 的大作中提到】 : With the advent of high density chip technology, most biological experiments : have been or will be performed in high-throughput. Tons of data are : generated every day across the world, which are required to be deposited into : some central repositories by most journals. For 70% human diseases, anyone : can retrieve thousands of data across a variety of measurements from public : repositories for free. There is a huge gap in between the data and : application in clinics. No matter you are a bioinformatician or not, as long : as you can make discovery from these data, validate them in new clinical : samples, and translate them into clinics, there is huge market for you. I : think that bioinformatics is transforming the biology and clinics. In the
I agree what demoner and AnthonyGe said are facts. One difference is what is the task of bioinformatics scientists. If it is just data analysis, I agree that it might expand for several years and then shrink. However, more and more bioinformatics scientists are directly proposing and investigating novel biological/clinical questions. Because we can easily retrieve ten times more data than any specialized biological lab on any disease, and buy samples/kits/mouse/tests online, we can very often answer the questions fast or answer questions which any specialized lab will never be able to ask. Our specialty is that we are from from interdisciplinary background and are willing to take risk and effort to enter any novel and exciting field.
b*1
54 楼
Data, samples, platforms, and even experiments are becoming commodity. You can easily buy all these online. A lot of them are even free. As long as you can propose important biological/clinical questions, you can get all process done without any single experiment by yourself. For example, with several gene/markers in hand and offer money and co-authorship for validation, any famous specialized lab would happily take the offer. These commodities freed us bioinformatics scientists from doing tedious experiments and focusing on finding and answering questions.
b*1
55 楼
I agree that there are still a lot of noise in the data due to the platform, and this is why some people on this board are cautious about bioinformatics . However, I believe that when you have hundreds and thousands of data, you can find the gold from the noise. The most important thing is the validation ! Don't stop at the prediction. It is our task to convince people that we can find clinically useful diagnostics biomarkers/drugs from these high- throughput data, so the world will follow. Because we bioinformatics scientists have interdisciplinary knowledge to propose biological/clinical questions, and the computational skills to identify and integrate data to answer questions, we can answer questions in the most efficient way. We are the best candidate to take advantage of the exploding commodity of data/samples/ experiments and lead the community efforts to move the biology ahead.
【在 d*****r 的大作中提到】 : I think the bottle-neck is not bioinfo itself, or statistical : methods/algorithm, or even math if we put aside the philosophical : question about whether biology can be understood by math. : The bottle-neck is high-throughput methods to acquire the data, : including : large-scale perturbing/imaging tools on molecular/cellular/animal level. : This requires huge progress on : instrumentation/nanotechnology/microfluidics/automation/material : science/surface chemistry, etc. That's why Feng Zhang was so popular on : the job market. It's also hard because this kind of tedious and non-
the problem is, the current dataset is still too small, covering only very very small portion of the disease or protein signaling dimension. Most of the data are 2D, or much less 3D, very very little 4D... e.g., we don't have any real-time molecular interaction data of hundreds protein or thousands of proteins in subcellular resolution in vitro, not even say in vivo... Without all these technic available, bioinfo can't go far.
platform, bioinformatics data, you validation we high- biological/clinical to answer the
【在 b*******1 的大作中提到】 : I agree that there are still a lot of noise in the data due to the platform, : and this is why some people on this board are cautious about bioinformatics : . However, I believe that when you have hundreds and thousands of data, you : can find the gold from the noise. The most important thing is the validation : ! Don't stop at the prediction. It is our task to convince people that we : can find clinically useful diagnostics biomarkers/drugs from these high- : throughput data, so the world will follow. Because we bioinformatics : scientists have interdisciplinary knowledge to propose biological/clinical : questions, and the computational skills to identify and integrate data to answer : questions, we can answer questions in the most efficient way. We are the
the current dataset is still too small.... bioinfo can't go far 人类基因组的原始数据有多大? 蛋白质组/代谢组的原始数据又有多大? bioinfo仅仅局限于分析蛋白结构么?井底之蛙。 了解蛋白功能,现有的数据都研究透了么?急切需要4D的数据?
【在 d*****r 的大作中提到】 : the problem is, the current dataset is still too small, covering only : very very small portion of the disease or protein signaling dimension. : Most of the data are 2D, or much less 3D, very very little 4D... : e.g., we don't have any real-time molecular interaction data of hundreds : protein or thousands of proteins in subcellular resolution in vitro, not : even say in vivo... : Without all these technic available, bioinfo can't go far. : : platform, : bioinformatics
【在 c****r 的大作中提到】 : the current dataset is still too small.... bioinfo can't go far : 人类基因组的原始数据有多大? : 蛋白质组/代谢组的原始数据又有多大? : bioinfo仅仅局限于分析蛋白结构么?井底之蛙。 : 了解蛋白功能,现有的数据都研究透了么?急切需要4D的数据?
I usually use Microsoft Word to do this kind of jobs. Open file -> Ctrl+H (Replace) -> Replace """ with "" A very good practice for all bioinfo guys is to learn Office as much as you can. Your neighbor's case is a very good example.
Memorizing all the hot keys (Ctrl + H in this case) also shows you are a professional data miner.
you
【在 K****n 的大作中提到】 : I usually use Microsoft Word to do this kind of jobs. : Open file -> Ctrl+H (Replace) -> Replace """ with "" : A very good practice for all bioinfo guys is to learn Office as much as you : can. Your neighbor's case is a very good example. : : variable
【在 K****n 的大作中提到】 : I usually use Microsoft Word to do this kind of jobs. : Open file -> Ctrl+H (Replace) -> Replace """ with "" : A very good practice for all bioinfo guys is to learn Office as much as you : can. Your neighbor's case is a very good example. : : variable
K*n
77 楼
depends on the format and your logic regular expression may help.
【在 e*****t 的大作中提到】 : 小小考你一下,如果同一个dataset,别的variable也有引号,但是你不想把他们也去 : 掉的话,怎么办?呵呵 : : you
b*1
78 楼
For industry jobs, biosta is narrower, focusing on analyzing clinical trial data. Bioinfo is more flexible. In terms of salary, biosta is pretty good overall. Bioinfo is more diversified. Data analyst has lower salary, but bioinfo scientists who understands biological/clinical questions has higher salary.
【在 K****n 的大作中提到】 : My browser does not decode the Chinese charactors of your club webpage... ( : UTF-8 encoding ??) Do I need to apply for membership?
K*n
84 楼
Thanks. My browser doesn't like it... will try again in a Chinese language env
I agree that we do not have enough quantitative data at dimension of the space/time to build a model to illustrate mechanisms of most disease. However, biology is still at the stage to fill all missing pieces. We do have quantitative data at hundreds of different molecular measurements to fill these missing pieces. For almost every human disease, I have thousands of measurements at mRNA, protein, genetics, genomic, cytokine, cellular levels in my database.
【在 d*****r 的大作中提到】 : the problem is, the current dataset is still too small, covering only : very very small portion of the disease or protein signaling dimension. : Most of the data are 2D, or much less 3D, very very little 4D... : e.g., we don't have any real-time molecular interaction data of hundreds : protein or thousands of proteins in subcellular resolution in vitro, not : even say in vivo... : Without all these technic available, bioinfo can't go far. : : platform, : bioinformatics
w*y
86 楼
呵呵, data analysis没你想的那么简单 基本上没有什么应用你对背后算法不懂, 按按按钮就能分析对 即使最简单最容易用的ncbi序列比对, 为什么要mask low complexity region, e value意义是什么, 很多人都不知道所以然 更不要说很多web service自己就是错的, 比如某公司的qPCR data analysis web service, 里面至少有2个明显的错误, 用它的biologists很多, 拿到的结果大多是错的misleading的, 只不过发发paper, 对错都无所谓而已
【在 K****n 的大作中提到】 : I usually use Microsoft Word to do this kind of jobs. : Open file -> Ctrl+H (Replace) -> Replace """ with "" : A very good practice for all bioinfo guys is to learn Office as much as you : can. Your neighbor's case is a very good example. : : variable
The problem of BME as a PhD program is the training is not systematic on any one direction. So in reality, most BME PhD students learned a little bit of almost every area of science and engineering. Although the problem solving skills are trained extensively in practice connecting any engineering tools with biology, I still think this connection should be done in the period of last 1-2 years of PhD, or postdoc, or even faculty period.
The reason is two folds: 1. recent progress of this area still rely so much on solid understanding of any one of the traditional science/engineering direction. For example, in tissue engineering, you can accredit most of the contributions to chemical engineering, or specifically, process engineering, surface chemistry, transport, materials, etc; it's the similar case in microfluidics area if you added in mechanical control electronics. 2. if you want to get into industry, the problem is more serious. Large companies only recruit fresh graduates with very specific skillset. This is why Merck would like to rather recruit chemical engineering or computer science PhD graduates to work on bio-related staff. In compannies, at the junior level, if you are not strong in a very specific skill set, it will be problematic. For startups, it would be a little different, but you still need to have a specific innovation to break into the market, not only just connecting the two area. As to "为数学/物理/计算机背景出身的学生所设立的生物课,大家都踊跃参加;反之 ,为生物背景 的学生设的简单数据分析+编程课,参加的人一次比一次少". The reason is simple, when you trained as math/physics/computer science, you will know exactly what you can contribute to this field after you learn some biology. If you failed after some practice, you can still go back and find a good job in companies, but not vise versa...biology is still very descriptive.
【在 d*****r 的大作中提到】 : The problem of BME as a PhD program is the training is not systematic on : any one direction. So in reality, most BME PhD students learned a : little bit of almost every area of science and engineering. Although : the problem solving skills are trained extensively in practice : connecting any engineering tools with biology, I still think this : connection should be done in the period of last 1-2 years of PhD, or : postdoc, or even faculty period. : : The reason is two folds: : 1. recent progress of this area still rely so much on solid
I agree BME could be a good undergrad major, but not a good PhD major. Biology is never a good undergrad major, but it also depends on which school, in some schools, I think biology major is trained in BME way. A good example is Misha Mahowald, she was a biology undergrad major in Caltech, but became one of the most innovative engineers in the world: http://www.witi.com/center/witimuseum/halloffame/1996/mmahowald.php Actually, her PhD thesis has led the creation of a whole new field and a new CNS department in Caltech. I think her work is at the deepest level of integrating biology and engineering, and the true "BME" in my heart. Hadn't she died too early, we would have some major progress on the fundamental principle of brain micro-circuitry and computation architecture.
i tried to tease you guys. u guys take it as a serious question.
serious.
【在 K****n 的大作中提到】 : She said there were only several thousand lines. Please don't be so serious. : .. are we really going to discuss this??
K*n
124 楼
never. but if I looked like serious to you, i won, haha
【在 e*****t 的大作中提到】 : i tried to tease you guys. u guys take it as a serious question. : : serious.
d*r
125 楼
I think that's because most of the bioinfo jobs are still academic- oriented or require extensive research experience. good or bad, that means bioinfo hasn't been commercialized enough as IT, and job duty was not divided enough.
are
【在 e*****t 的大作中提到】 : the situation is ms in biostat get hard time to find a job, while phds are : okay.
d*r
126 楼
right, I think you are the most non-serious one, :)
【在 K****n 的大作中提到】 : never. but if I looked like serious to you, i won, haha
Bioinfo is way more than statistical analysis/software-development and is a broad discipline(structure,function,pathogenesis...) Goal is to consolidate, correlate data & knowledge in a quantitative manner( physicical/chemical principles,math models) for improved understanding and new hypothesis generation. It is a critial component of bio nowadays, because given the data complexity and diversity, manual operation and low- dimension(e.g. pair-wise) interpretation is no longer effective. Both bioinfo and bench evolve together via hypothesis-experiment test-new hypothesis iterations. no point to argue who is superior who is doomed. NGS opens a new door to "omics" era, it can do and do better all what array can(genomic/transcriptomic/epigenomic), and go beyond. e.g. handle degraded samples,sense-explicit profiling, digital yeast-2-hybrid... Our dept has a hiseq2000 and there are just so many exciting projects we can do now. Btw, anyone looking for NGS-oriented bioinfo research position(phd level) in pharma can drop me a note. we have an opening.
~~~~~~~~~~~~~~~~~~~~~~ this is an important question and a big question mark so far. faithful or wishful thinking does not help answer it. for the modeling part, ph.d. level knowledge would be necessary for the information processing part, there are many CS/CE masters who can do the job real quick...
你是在 Industry 做的,在你找工作的时候,有没有感觉到工业界生物信息的工作还是 比其他行业的平均水平少?或者是比academy 的生物信息的工作少很多?
a broad manner( array
【在 y********a 的大作中提到】 : Bioinfo is way more than statistical analysis/software-development and is a broad : discipline(structure,function,pathogenesis...) : Goal is to consolidate, correlate data & knowledge in a quantitative manner( : physicical/chemical principles,math models) for improved understanding and : new hypothesis generation. It is a critial component of bio nowadays, : because given the data complexity and diversity, manual operation and low- : dimension(e.g. pair-wise) interpretation is no longer effective. : Both bioinfo and bench evolve together via hypothesis-experiment test-new : hypothesis iterations. no point to argue who is superior who is doomed. : NGS opens a new door to "omics" era, it can do and do better all what array
If you have a few candidate genes with convincing stories, and offer co- authorship and experiment cost, although every lab would love to validate your findings. One key is that you need to find a simple and effective experiment for validation. Now days, tons of experiments have become commodities.
【在 o********r 的大作中提到】 : GO, pathway之类的说实话没太大实际意义,因为无论GO或者pathway数据库中都有不少 : 的错误和遗漏。NGS倒是可以直接发现一些direct AA change以及一些fusion protein : ,可能更有直接意义,不过这个需要deep coverage。 : : list
s*l
201 楼
So many discussions here, this point I think is the most accurate one. All the new techs generate more data, higher quality than before,new types of data (just assume some new tech will do these in the future), the most important step is how to get insights or generate potential predictive things in biology. Developing new ways to interpret and making sense of these complex data is always critical.