bioinformatics新手指导

“What topics should I study to learn bioinformatics?” – I often get asked this question by biology students and sometimes even biology professors. The answer depends on what you want to do, but how would someone, who is new to using computers for biology, know what can be done? To add to our difficulties, technology is changing so rapidly that even experts do not know what software tools will remain useful next year.

Keeping the above complications in mind, I created a general framework to describe one’s abilities in bioinformatics that should be valid even with changing technologies. Another reason for creating this framework is to fit our new posts into one or other category for easy description.

From a top level, I prefer to divide the levels of expertise in bioinformatics into five layers with 5 being the most difficult.

Layer 1 – Using web to analyze biological data

Layer 2 – Ability to install and run new programs

Layer 3 – Writing own scripts for analysis in PERL, python or R

Layer 4 – High level coding in C/C++/Java for implementing existing algorithms or modifying existing codes for new functionality

Layer 5 – Thinking mathematically, developing own algorithms and implementing in C/C++/Java

Layer 1 – Using web to analyze biological data

By now, almost all biologists are familiar with using the web to run BLAST on a new sequence at NCBI, or to fold a short sequence online and check whether it is a miRNA, or to book hotel rooms for a conference. Ok, the third example is not bioinformatics. I was only checking whether you can tell the difference, and if you did, you already got an A+ in Layer 1

I was initially reluctant to split this into a different category, because using the web to find information seems so easy !! However, the main purpose of separating this is to show respect to the programmers, who develop online interfaces. Ease of using the web tends to make biologists think that the web interfaces are easy to develop. So, they allocate $50 on a $1 million grant for presenting their data, and then get surprised, when their website crashes within few weeks of publication. In reality, developing high quality interfaces is very skill intensive, but also has good return in terms of visibility of one’s work. As a rule of thumb, easier an website is to use, more difficult it is to develop. Skilled web programmers prefer to work for Facebook or Zynga for $$$ rather than wasting time with researchers, who show no respect.

Layer 2 – Ability to install and run new programs

Usually, most cutting edge programs (Bowtie, Velvet, Trinity, Stampy, etc.) do not come with an web interface, because the developers neither have time nor computing resources to provide web services for everyone. Moreover, even the traditional programs like BLAST or Interpro-scan cannot be run for thousands of genes through web interfaces, because data-intensive web services cost mucho dinero. It is like using someone else’s kitchen for free to run a full-service restaurant – never gonna happen. Therefore, sooner or later, each user will need to learn downloading and running programs to stay on the cutting edge of biology.

What skills and technologies are required for layer 2? Firstly, you need your own computer, and what kind of computer to get is a major decision. “Should it run Windows, Mac or Linux?” “How much RAM should it have?” “How much hard drive should it have?” “Should we buy one machine or a cluster?” “Should we just use high-performance computing facility at the university or Amazon cloud?” Those are all important questions that need to be asked, if one wants to fully utilize capabilities of next-gen sequencing machines. Let us address the issue about computers in a later post and restrict ourselves only to computational skills.

The good news is that most programs these days can be easily installed in Windows, Mac or Linux. I remember that in earlier day a program developed for one type of Unix had to be almost rewritten for another Unix variety.

Regarding computational skills for this layer, one needs to learn how to install and run programs from the command line. These days, people see computers as pictures, singers or dancing beauties, and not someone speaking strange short words such as ‘ls’, ‘mv’, ‘cp’, ‘rm’ that, if misspelled, can erase your last five years’ work in 5 seconds. However, every computer can still speak those sweet words, if treated nicely. The challenge of layer 2 is to learn that language of computers.

The best way to master layer 2 is to first get familiar with UNIX or Windows command line, and then ask for an expert’s help in installing and running various programs. Carefully watch and write down every word he is typing, and ask for what they mean in details. You will get a hang of it after few rounds of trial. Apart from learning how to use UNIX, there is no course to take to master layer 2, because every program is different installation and running. Just like golf, practice, practice and practice. For help, online forums for specific programs are very useful resources.

Layer 3 – Writing own scripts for analysis in PERL, python and R

Let us start with an example. A biologist has large volume of expression data and wants to find expression levels of 300 transcription factors that he likes. If he had only three genes, he could have cut and pasted the lines. Cutting and pasting for 300 genes become a bit laborious and time consuming. As another example, let us say a biologist wants to extract coordinates 7000-8500 from scaffold_1 of a 300M base genome. This task is also technically possible by loading the genome into a word processor, but it is not the most desirable method. Asking a highly skilled programmer to do the above task becomes sub-optimal use of resources, because the skilled programmers tend to prefer layers 4 or 5 (explained below) for glory.

Learning a bit of PERL or python can do wonders for small tasks described above and should be part of every biologist’s toolbox. R is a statistical software that plays similar role for statistical analysis of data, when Excel is no longer feasible.

The reason I placed this above layer 2 is because layer 3 needs some conceptual or mathematical thinking in designing the programs and therefore is not always straightforward. PERL, python and R can do wonders, when applied by very skilled programmers. For example, the entire gbrowse genome browser is written in PERL. On the other hand, for scripting languages can also be applied by beginners to save time on manual tasks.

In not too distant future, Layers 1-3 will be needed by almost every biologist, whereas layers 4 and 5 will remain the domains of experts. Therefore, I will cover those two layers in a new post.

Layer 4 – High level coding in C/C++, Java for implementing existing algorithms or modifying existing codes for new functionality

Those who try to be on the cutting edge of solving biological problems often reach the limits of layers 1-3 very quickly. This is even more important these days, because biology is getting very data intensive. Let me give few examples -

i) Let’s say a lab invested money for an ABI SOLiD sequencing machine, but a new paper on aligning sequences came out that does not cover SOLiD color-space data yet. Should the lab throw away its sequencing instrument and buy something new?

ii) As another example, let’s say a lab has arrangements with a high-end computing center to use 100 parallel processors, but the best assembly program he found only runs on single machine that requires 512 Gb of RAM. Should the lab buy another new RAM-heavy computer ($50-100K) ?

iii) A supposedly very accurate algorithm was described in a journal, but the authors only provided assembled code and not source code for their method. You have a different data type (let’s say SOLiD data or shorter read lengths than the algorithm described). What should you do?

Various questions of above type come up almost every day, and modifying existing software codes is the best way to address them. This layer requires expertise in one or more high level programming languages such as C, C++ and Java, and ability to understand algorithms described in the journal. It is absolutely necessary, if a a lab or core facility wants to fully utilize its capabilities. By restricting itself only to readily developed programs, the lab is unable to fully utilize the capabilities of its expensive instruments.

Layer 5 – Thinking mathematically, developing own algorithms and implementing in C/C++/Java codes

A new efficient algorithm gets used millions or billions of times over many years by labs all around the world. BLAST is one good example. To provide some recent examples on next-gen sequencing, alignment programs such as Bowtie and BWA used Burrows-Wheeler transform, a mathematical transform, to dramatically increase speed of aligning many short sequences on a reference genome. Next-gen assembly programs such as Velvet or SOAPdenovo use de Bruijin graphs, a graph-theoretic technique, and can allow assembly of large genomes for a fraction of costs of even five years back.

Developing these programs require mathematical brain, which is far from biology. One of my motivations for this blog is to explain some of these algorithms in simple languages so that those intrepid souls, who like to experiment on new codes, can try them out easily.

From：http://www.homolog.us/blogs/2011/07/22/a-beginners-guide-to-bioinformatics-part-i/

http://www.homolog.us/blogs/2011/07/22/a-beginners-guide-to-bioinformatics-part-ii/