Hufeng Zhou's Blog

Tuesday, 30 September 2014

15條秘訣，讓你今天就有高效生產力！

1. 比平常早起一小時。
2. 在周日晚上制定好下周的計畫。
3. 承認自己能力有限，學會拒絕。忽略瑣碎的事，關注於想實現的目標。
4. 關掉你手機上的電子郵件通知功能，以免不斷干擾。
5. 使用簡單的待辦事項列表工具，每天在列表中只安排3件一定要做的事情。
6. 根據最重要的重大任務安排每天的計畫，而不是從最簡單的小任務開始。
7. 盡量減少避免開會，如果真的必須開會，務必要得到具體的成效。
8. 每次的談話結束後，都知道接下來該怎麼做。
9. 每工作2小時就做20分鐘內的適當休息，如聆聽喜歡的歌曲、出去走一走。
10. 真心問自己「我現在做的事，是不是能讓我更接近自己想實現的目標？」
11. 在筆記本上記下讓自己分心的事情，一個星期後回來檢視。
12. 別再拖了，現在就開始行動！
13. 停止上網，如果你正在從事不需要上網的工作，就把瀏覽器關掉。
14. 別再說「我應該做」，而是「我會做」。想想尤達大師曾說：「沒有試試看，只有做與不做。」
15. 有能力分辨重要與不重要的事情。

Saturday, 27 September 2014

Why is pre-packaged ground beef red on the outside and sometimes dull, grayish-brown inside?

I'll quote the full-explanation from the USDA, just for reference:

Oxygen from the air reacts with meat pigments to form a bright red color which is usually seen on the surface of meat purchased in the supermarket. The pigment responsible for the red color in meat is oxymyoglobin, a substance found in all warm-blooded animals. Fresh cut meat is purplish in color. The interior of the meat may be grayish brown due to lack of oxygen; however, if all the meat in the package has turned gray or brown, it may be beginning to spoil.

Friday, 26 September 2014

Do the important but not urgent things in life.

This popular book, Scarcity: Why Having Too Little Means So Much, uses new studies and new theories to explain an old classic finding. Doing only the urgent things in life, will have long term consequences in life, leaving one no time nor resource to do the important but not very urgent things. People who are constrained with time and money will fell in to the mindset of habitually doing the most urgent things very easily, eventually wasted valuable life on urgent but not important things and die poor and unhappy.

Do the important but not urgent things in life is the key to success.

Wednesday, 24 September 2014

Restriction sites in your sequences

A question we get asked all the time–and I see on mailings lists with about a twice/year frequency–is where can I look at the restriction sites in my sequences? I’m going to mention a few that I know about, but if anyone has other favorite tools or spots, do let us know in the comments.
UCSC Genome Browser: if you look in the browser in the “Mapping and Sequencing” Track groups area there is a menu called Restr Enzymes. I have turned this track on, and turned off some other default tracks, and created a “Session” to show you exactly how this can look. I picked a small gene for the example. Load my session: http://genome.cse.ucsc.edu/cgi-bin/hgTracks?hgS_doOtherUser=submit&hgS_otherUserName=Mary%20Mangan&hgS_otherUserSessionName=restriction_enzymes
You can see the sites in Ensembl, too. From that UCSC session I provided you can click the link in the blue navigation area at the top of the UCSC page to go to the exact same region in Ensembl. In the “Detailed View” section at Ensembl you can see menus for “Decorations” and other display features. In Decorations you can check the box for Rest. Enzymes to see them on your viewer.

However, there may be times where your sequences is different from genomic, or not available in a species browser, and sometimes you just want another type of viewer. You might try PlasMapper (http://wishart.biology.ualberta.ca/PlasMapper). I took the same sequence from the browsers (the Casp-1 gene) and put the FASTA sequence in, and created a really quick map. It also gives text output if you prefer. It can add a display for some handy features that you might have on plasmids or constructs like promoters or affinity tags. Takes only a few seconds, a very handy little freely available tool for you to know about. I know it isn’t new, but it is such a common question for us that we thought we would mention it.
Dong, X., Stothard, P., Forsythe, I.J., Wishart, D.S. (2004). PlasMapper: a web server for drawing and auto-annotating plasmid maps. Nucleic Acids Research, 32(Web Server), W660-W664. DOI: 10.1093/nar/gkh410

Friday, 19 September 2014

Google HR talk about Resume

I've sent out hundreds of resumes over my career, applying for just about every kind of job. I've personally reviewed more than 20,000 resumes. And at Google we sometimes get more than 50,000 resumes in a single week.

I have seen A LOT of resumes. Some are brilliant, most are just ok, many are disasters. The toughest part is that for 15 years, I've continued to see the same mistakes made again and again by candidates, any one of which can eliminate them from consideration for a job. What's most depressing is that I can tell from the resumes that many of these are good, even great, people. But in a fiercely competitive labor market, hiring managers don't need to compromise on quality. All it takes is one small mistake and a manager will reject an otherwise interesting candidate.
I know this is well-worn ground on LinkedIn, but I'm starting here because -- I promise you -- more than half of you have at least one of these mistakes on your resume. And I'd much rather see folks win jobs than get passed over.
In the interest of helping more candidates make it past that first resume screen, here are the five biggest mistakes I see on resumes.
Mistake 1: Typos. This one seems obvious, but it happens again and again. A 2013 CareerBuilder survey found that 58% of resumes have typos.
In fact, people who tweak their resumes the most carefully can be especially vulnerable to this kind of error, because they often result from going back again and again to fine tune your resume just one last time. And in doing so, a subject and verb suddenly don't match up, or a period is left in the wrong place, or a set of dates gets knocked out of alignment. I see this in MBA resumes all the time. Typos are deadly because employers interpret them as a lack of detail-orientation, as a failure to care about quality. The fix?
Read your resume from bottom to top: reversing the normal order helps you focus on each line in isolation. Or have someone else proofread closely for you.
Mistake 2: Length. A good rule of thumb is one page of resume for every ten years of work experience. Hard to fit it all in, right? But a three or four or ten page resume simply won't get read closely. As Blaise Pascal wrote, "I would have written you a shorter letter, but I did not have the time." A crisp, focused resume demonstrates an ability to synthesize, prioritize, and convey the most important information about you. Think about it this way: the sole purpose of a resume is to get you an interview. That's it. It's not to convince a hiring manager to say "yes" to you (that's what the interview is for) or to tell your life's story (that's what a patient spouse is for). Your resume is a tool that gets you to that first interview. Once you're in the room, the resume doesn't matter much. So cut back your resume. It's too long.
Mistake 3: Formatting. Unless you're applying for a job such as a designer or artist, your focus should be on making your resume clean and legible. At least ten point font. At least half-inch margins. White paper, black ink. Consistent spacing between lines, columns aligned, your name and contact information on every page. If you can, look at it in both Google Docs and Word, and then attach it to an email and open it as a preview. Formatting can get garbled when moving across platforms. Saving it as a PDF is a good way to go.
Mistake 4: Confidential information. I once received a resume from an applicant working at a top-three consulting firm. This firm had a strict confidentiality policy: client names were never to be shared. On the resume, the candidate wrote: "Consulted to a major software company in Redmond, Washington." Rejected! There's an inherent conflict between your employer's needs (keep business secrets confidential) and your needs (show how awesome I am so I can get a better job). So candidates often find ways to honor the letter of their confidentiality agreements but not the spirit. It's a mistake. While this candidate didn't mention Microsoft specifically, any reviewer knew that's what he meant. In a very rough audit, we found that at least 5-10% of resumes reveal confidential information. Which tells me, as an employer, that I should never hire those candidates ... unless I want my own trade secrets emailed to my competitors.
The New York Times test is helpful here: if you wouldn't want to see it on the home page of the NYT with your name attached (or if your boss wouldn't!), don't put it on your resume.
Mistake 5: Lies. This breaks my heart. Putting a lie on your resume is never, ever, ever, worth it. Everyone, up to and including CEOs, get fired for this. (Google "CEO fired for lying on resume" and see.) People lie about their degrees (three credits shy of a college degree is not a degree), GPAs (I've seen hundreds of people "accidentally" round their GPAs up, but never have I seen one accidentally rounded down -- never), and where they went to school (sorry, but employers don't view a degree granted online for "life experience" as the same as UCLA or Seton Hall). People lie about how long they were at companies, how big their teams were, and their sales results, always goofing in their favor.
There are three big problems with lying: (1) You can easily get busted. The Internet, reference checks, and people who worked at your company in the past can all reveal your fraud. (2) Lies follow you forever. Fib on your resume and 15 years later get a big promotion and are discovered? Fired. And try explaining that in your next interview. (3) Our Moms taught us better. Seriously.
So this is how to mess up your resume. Don't do it! Hiring managers are looking for the best people they can find, but the majority of us all but guarantee that we'll get rejected.
The good news is that -- precisely because most resumes have these kinds of mistakes -- avoiding them makes you stand out.
In a future post, I'll expand beyond what not to do, and cover the things you should be doing to make your resume stand out from the stack.

Thursday, 18 September 2014

Whome Exome Sequencing

PREPROCESS:

Index human genome (Picard), we used HG18 from UCSC.
Convert Illumina reads to Fastq format
Convert Illumina 1.6 read quality scores to standard Sanger scores

FOR EACH SAMPLE:

Align samples to genome (BWA), generates SAI files.
Convert SAI to SAM (BWA)
Convert SAM to BAM binary format (SAM Tools)
Sort BAM (SAM Tools)
Index BAM (SAM Tools)
Identify target regions for realignment (Genome Analysis Toolkit)
Realign BAM to get better Indel calling (Genome Analysis Toolkit)
Reindex the realigned BAM (SAM Tools)
Call Indels (Genome Analysis Toolkit)
Call SNPs (Genome Analysis Toolkit)
View aligned reads in BAM/BAI (Integrated Genome Viewer)

SAMPLE SCRIPT

/bin/BWA/bwa aln -f /output/FOO.sai -t 3 /seq/REFERENCE/human_18.fasta /seq/FQ/FOO.sanger.fq
/bin/BWA/bwa samse -f /output/FOO.sam /seq/REFERENCE/human_18.fasta /output/FOO.sai /seq/FQ/FOO.sanger.fq
/bin/SAMTOOLS/samtools import /seq/REFERENCE/human_18.fasta /output/FOO.sam /output/FOO.bam
/bin/SAMTOOLS/samtools sort /output/FOO.bam /output/FOO.sorted
/bin/SAMTOOLS/samtools index /output/FOO.sorted.bam /output/FOO.sorted.bam.bai
java -jar /bin/GTK/GenomeAnalysisTK.jar -T RealignerTargetCreator -R /seq/REFERENCE/human_18.fasta -I /output/FOO.sorted.bam  -o /output/FOO.intervals
java -jar /bin/GTK/GenomeAnalysisTK.jar -T IndelRealigner -R /seq/REFERENCE/human_18.fasta -I /output/FOO.sorted.bam -targetIntervals /output/FOO.intervals --output /output/FOO.sorted.realigned.bam
/bin/SAMTOOLS/samtools index /output/FOO.sorted.realigned.bam /output/FOO.sorted.realigned.bam.bai
java -jar /bin/GTK/GenomeAnalysisTK.jar -T IndelGenotyperV2 -R /seq/REFERENCE/human_18.fasta -I /output/FOO.sorted.realigned.bam -O /output/FOO_indel.txt --verbose -o /output/FOO_indel_statistics.txt
java -jar /bin/GTK/GenomeAnalysisTK.jar -T UnifiedGenotyper -R /seq/REFERENCE/human_18.fasta -I /output/FOO.sorted.realigned.bam -varout /output/FOO.geli.calls -vf GELI -stand_call_conf 30.0 -stand_emit_conf 10.0 -pl SOLEXA

 Between step 7 and 8, it is recommended to add an additional step:



samtools calmd -Abr FOO.sorted.bam human_18.fasta > FOO.baq.bam

This will substantially improve SNP specificity.



A few other comments to David Quigley's pipeline are:




It is recommended to use Dindel
 for indel calling. It is the few indel callers that properly model 
indels (most not, including IndelGenotyperV2). Broad is reimplementing 
the Dindel model in GATK. The latest samtools also models indels properly, but it is not evaluated as widely as Dindel.
For the human genome build 36, it is recommended to use the version used by the 1000 genomes project. You will lose genes in the pseudoautosomal regions if you use hg18. Hg19/b37 is fine.
The samtools
 and GATK SNP callers are converging in algorithms and in results. GATK 
is arguably better than MAQ, which has been confirmed by several 
independent studies.