Jul 012015

Sometimes when things occupy my mind, I like to write it down. If I don’t mind sharing it with others, I will put this on my blog. However, when I have manuscripts or grants due (actually, when do we ever not have ms or grants due?), I often feel guilty writing blogs (didn’t you have more important work to do?).

Recently I talked to Terry Speed about this. He thought that writing blog could be a healthy activity if: 1) I can get the thoughts out of my mind so I can focus on other important issues; 2) if I find writing blogs relaxing (can’t always be so stretched with writing grants and papers); and 3) if the blogs could be beneficial to other readers. I guess I can consider writing blog like doing exercise or reading for fun, and don’t need to feel guilty for writing blogs. Writing is especially helpful when I travel. Sometimes sitting in the plane / train / taxi when it is a bit bumpy, I have a lot in my mind yet I don’t have wireless internet, or if I get lost in a conference / seminar talk but don’t feel like focused enough to revise a paper or write a grant, I can type my thoughts without looking closely about what I write, and it feels good to get my thoughts filed away.

Jun 242015

For years, I have been trying to have regular physical exercise without much success. I tried to run or swim, but between the unpredictable weather, my travel schedules, and monthly cycles, it is hard to keep a daily practice. I also tried the 7MWC. It is fast, but so hard and painful it is challenging for me to keep up.

I took some Yoga class during graduate days, but for some reason never liked it. For many years, I considered Yoga not real exercise. Recently I started doing YogaStudio from my iPhone, since it had amazing reviews. It has 15min, 30min, or 60min readymade classes, focusing on different things at different levels. This is something I can reliably do daily either in the morning or at night (actually nice to do right before bed). At the beginning, my mind kept on wandering around, thinking about some research problems, family errands, the need to buy something or write an email. But recently I did the Yoga studio with a friend, and noticed her smooth and deep breathing. I tried to mimic her, and realized that the long and even breathing naturally keeps my mind focused and relaxed. It has been great, and I now bring my yoga mat even when I travel.

Another colleague mentioned that it is important to have some daily exercises that will get our heart rate to > 150 beats / min. I started practicing climbing stars, as it definitely jumps by heart rate. I often park at P4 and my office is at the 11th floor, so going up and down 15 floors is becoming a routine. I read from a book “Better Than Before” that the benefit of exercise doesn’t have to come from rigorous exercise, rather the basic movement of walking, stretching, wiggling, going up stairs could be most beneficial. This is definitely encouraging, so I am trying to keep these simple things when I have a min or so.

P.S. Learning that a lot of the successful colleagues that I admire, e.g. Bing Ren, Terry Speed, Peter Jones, routinely run, I am trying to give running another try. I am surprised to find that after my recent stair exercises, getting back into running is not that painful.

Jun 102015

Recently I noticed something. For a popular software tool that handles big data, like RMA, BWA, DESeq, or MACS2, the community usually gives it enough credit, because users often cite these tools in their studies. However, for web database resources that people can quickly look up, they are often extremely widely used and yet not get enough credit from citation. Examples include: UCSC genome browser, CBioPortal, and BioGPS. They provide user-friendly visualization of public data, so people can refine their hypothesis or check things quickly. Unfortunately these tools (unless bundled with another biological study), rarely get published well beyond NAR. However, in terms of their value to the scientific community, how are they not as good as a Cell Resource paper? I hope the community will consider the number of visits / queries in evaluating the value of such web resources, instead of their published journals and citations by others.

May 102015

Recently I have been a bit overwhelmed with work load, which reminded me of the “work fixed number of hours and in fixed amounts” recommended in this blog. I decided to count up the numbers of days I spend each year on various tasks, and to my horror that quickly added up to more than 365 days / year! No wonder I am sleep deprived, stressed and cranky all the time. So I did some soul searching and considered which areas I could scale back in order to fit in 264 days / year (5.5 days / week * 48 weeks / year, which leaves 2 weeks of vacation / sick or personal days and 2 weeks of public holidays). This really gets me to re-think my schedule and decide what to cut out from my schedule.

Here is how the new schedule looks like:

    90 days (3 hours a day * 5 days a week * 48 weeks a year / 8 hours a day) on research project meetings (lab meeting, journal club, project update) within my group, meeting with collaborators or seminar speakers, and going to local research seminars.
    12 days (an afternoon twice a month) on reading papers and thinking about research ideas. As a scientist, I would really hope to at least double this time…
    40 days on writing / revision papers I am the corresponding authors on. Currently the papers on the pipeline include (in terms of stages): Han/Ted (SSC), Zhou/Yiwen (ceRNA), Xiaoqi/Naiqian (drug network), Tao (MACS2), Qing/Chengyang (EgLN2), George/Tao (MANCIE), Bo (immunology), Wei/Johannes (MAGeCK+VISPR), Haiyun/Yin (drug combination), Mulin (emsQTL), Alvin (ChiLin + Gnosis), Wenyu/Shunmin (Sox2), Peng (CRISPR predict), Su/George/Cliff (MARGE), Ted/Wei/Han (breast cancer CRISPR screen), Teng/Yiwen (RBP). That leaves ~3 days / paper (in fact most take > 5 days), including the initial submission and later revision. If a paper needs a lot of additional work, I am afraid I will have to rely on the first authors to spend more efforts revising it. In the future, I should try to reduce the total number of papers and improve the quality of our own publications.
    4 days on papers I am not the corresponding authors. Ideally I should reduce this time to 0 days and move the time to my own papers, and rely on the lab members who are on these projects to help with paper revision.
    36 days on teaching and talks. Every year I give about 24 lectures for the course I teach (http://stat115.org and HarvardX) at Harvard, which takes on average one day / lecture, in learning, preparing the lecture slides, planning the computer lab, HW with TA, and writing the finals. The rest of the talks includes teaching at Tongji, Longwood Translational Medicine Initiative, conference talks, and research seminars (including the day at the seminar), which takes ~ 1 day / month.
    20 days on writing grants, managing finances, preparing progress reports, and going to consortium meetings. In a soft money environment, it takes ~3 ongoing PI grants and a number of smaller grants (sub-contract, internal funding) to keep a lab going. We make ~3 major grant submissions / year, and it takes a full week to write a PI grant even with postdocs’ help to write different sections. I will have to rely more on lab members to share some of these responsibilities, encourage them to apply for fellowships, and minimize collaborative grants with < $33K / year (3% of me and 30% of a postdoc) budgets.
    17 days of grant and paper reviews. I will become the chair at GCAT this fall, which means 3 sessions * 5 days (~3 days of reviews + 2 days of meetings). I review ~8-9 grants / session (less once I become the chair, but I need to read all the discussed proposals ahead of time), which means I could only spend ~3 hours reviewing each grant (applicants really need to make their proposals easier to read / understand for reviewers). That only leaves 2 days / year for reviewing papers or dealing with editorial duties, so I will need for rely on lab members to review more papers with me (< 1 paper / month).
    20 days on conferences + travel. I tried to limit my travel to <= once a month, and some conferences last 3-5 days. This also includes the time organizing conferences (CSHA, GRC, IBW, YBPIW), such as speaker invitation and securing conference sponsorship (e.g. $100K for the 2017 Gordon conference!). I need to be more selective in agreeing to organize conferences in the future.
    6 days of committee (executive committee of research, faculty career / promotion) and faculty meetings, as well as SAB (PICB, CPRIT) meetings. I could only allocate 1/2 day / month, so will be resigning from some less important committees in the futures.
    10 days on letter writing for students / postdocs applying for awards / fellowships / jobs and faculty applying for new jobs and promotions. This also includes support letters for colleagues’ grant applications, people applying for green card, or invitations letters for visitors to my lab, which thankfully my assistant can mostly take care of. This means, no more than one promotion evaluation letter a month, and I will only write a letter if I already know their work (no time to specifically read their papers), and no more than one job / fellowship letter for students and postdocs a month. Also I will need at least two-month notice.
    4 days on lab member career advice and annual evaluations, and also on collaborators’ students / postdocs careers. I value the annual evaluation, which is quality time for postdocs / students to brainstorm and prepare for their future. This would also include postdoc rehearsing their job talks, teaching postdocs about how to write grants, give talks, apply for jobs, etc. This means I could only care about 16 junior people and spend 2 hours / year on each!
    4 days on retreats. For sure I need to go to my own lab retreat and our CFCE center retreat. Then I will need to select 2 days / year to go to DFCI and HSPH biostats retreat, Genetics (I am affiliated with Harvard BBS) retreat, epigenetics retreat, SPORE retreat, Broad retreat, colleagues’ promotion / retirement / birthday symposiums.
    1 day on visitors, including short term visitors, postdoc interviews, rotation students. I need to rely more on my lab members to host those people who just want to stop by, talk to me, or see Harvard / my lab, or just ignore these requests completely, so I can save time for real collaborators coming to town and truly promising postdoc and student candidates.

This is counting 8 hour / day, because I spend another 2-3 hours / day on emails and blogs, and another > 1.5 hours / day on commute (so I try to work at home more), etc!! Therefore, if I have spent less time on a paper revision, course, grant review, collaborative project, recommendation letter, conference organization, etc, or turn down a request, I hope my colleagues could forgive me. There are only so many hours one has and so many things one can do in a year. Also going through this schedule, now I appreciated a lot more any senior people or mentors who spared their time before in helping me, writing letters, giving me career advices, finding resources for my projects, etc, and will respect their time better in the future.

May 072015

I saw this blog article I wrote a long time ago but never posted. It has some interesting ideas, although over the years the graduate applicant pool at Tongji has changed so much that the following might not be applicable. However, for other bioinformatics programs about to screen their candidates and aspiring bioinformatics students preparing for graduate interviews, this blog might still be useful.

Since my collaboration with Tongji University in China in 2009, I have been involved in recruiting and screening graduate applicants to Tongji’s Bioinformatics graduate program. Most of these applicants just finished their junior year in college. They have very diverse backgrounds, and unlike graduate applicants in the US, often have little bioinformatics experience.

The first level of screening checks whether the CV appears professional, and some simple facts: current university (985/211 universities preferred), majors (maths, physics, biology, computer science, and bioinformatics preferred), ranking (top 10% GPA, and good scores on physics, maths, analytical chemistry, computer science and English courses preferred), and English test scores (CET6 > 500 preferred).

After the initial screen, I get in touch with students to talk for about ~30min over the phone or Skype (video conference is even better). Over the years, I have settled on the following questions. In addition to testing students’ motivation and maturation, this conversation also gives me some clues about the students’ logic thinking, professionalism and communication skills.

  • Can you tell me about your background? Current college, major, and research experience. Research experience is important but not necessarily (and probably better not) in bioinformatics.
  • Why are you interested in bioinformatics at Tongji University? Do they know what bioinformatics really is? Which institutions and scientists do good bioinformatics research? Do they know what PIs at Tongji Bioinformatics program do?
  • What are your goals for the coming 1-2 years, and after graduate school? Students with well thought out and concrete goals are self motivated, focused, and more mature. In addition to career goals, one might also consider self improvement, health, and personal happiness goals.
  • What is the biggest failure or setback in your life and what you have learned from this? Mental resilience is important for scientific research. It is important to challenge oneself, fail enough times, and learn from these experiences. If students tell me their lives have been pretty smooth with no major setbacks, I would ask why didn’t they try harder to get more of what they want in life.
  • What are your hobbies or extra-curriculum activities? Michael Zhang once told me, “if you don’t know how to play well, you don’t know how to do good research”. We want our students to bring an interesting aspect to our graduate program. I also like students who are socially conscious and responsible.

If a group of students are all on site (e.g. during their graduate interview weekend), I also like to get the students to play the Pictionary Game. We divide the students into teams of 4-5 people (one group is OK too). Each team ask one representative to come to the front, pick one card with an English word, then go back to draw it for the teammates to guess the word. The people drawing can only tell their teammates whether the word is none/verb/adj/adv and they can provide hint by nodding or shaking their heads but can not speak. Here we use abstract words that are hard to draw, such as create, try, and succeed. It is interesting to see how the people drawing convert abstract ideas into drawing, and which teammates can guess things nonstop until they succeed. It also test the students’ English vocabulary, IQ, social behavior, and team spirit.

Of course, the best test is for the student to spend a few weeks in the summer to get some bioinformatics training at Tongji. During their stay, the students will read and present some scientific papers, analyze some genomics datasets using available algorithms, and learn to program in python. Both the research group and the applicants get to know each other better. The students also meet individually with admission committee composed of senior graduate students, postdocs, and faculty who will provide their evaluation of the students later.

Apr 292015

Happened to see this article about Neal Copeland and Nancy Jenkins. Fascinating experiences!

I love their idea about selecting postdocs through the “lunch test”, and on postdocs with “50% success rate” to those “luck happens with hard work”. I should follow their advices when selecting students and postdocs in the future… It is a fun read. I only feel a little sad that they decided not to have kids. We are lucky that academic environment nowadays are much more friendly to female scientists with family. Also, there are enough successful female scientists now to form a support network.

Thinking again about this “lunch test”, you need to be smart, fun, and nice to do well here. I am reminded of my colleagues like Myles Brown, Soumya Raychaudhuri, Zhiping Weng, and Yu Xue. I am not so good in conference or consortium meeting meals. Some of the difficulty is due to race, gender, and confidence. In addition, I should have more fun in life, not necessarily in the form of having lots of fun activities like travel or sports, but genuine peace and joy in the heart. This would mean that we work hard only on the things we enjoy and drop the balls in some “academic niceties”. Hmm…

Apr 282015

Very often a computational genomics project in the lab requires the output of some tables. I noticed in many cases, some number columns have numbers that stretch VERY LOOOOOOONG. To display a chromosome location, it is totally legitimate to use a long number, e.g. chr1 42392716, because it is different from chr1 42392715 and chr1 42392717. However, in some other calculations, such as false discovery rate, transcription factor motif matching score, differential expression fold-change, it would not make sense to have 10 significant digits. For FDR, most people just want to know whether the gene / peak is around 1%, 5%, 20% or 80% FDR, so there is no need to show FDR as 1.238346226253182% (our FDR estimate simply doesn’t have that level of accuracy). Sometimes from cross validation, we know that the area under the curve of our prediction is around 0.813 (pretty good), then it would make sense to only use 2-3 significant digits to make classification predictions, even though the regression model might give 100 digits after the decimal points. Using fewer significant digits is often easier to visualize, creates smaller-sized files, and makes better sense.

Apr 042015

I have a long commute everyday. In order to stay awake during driving, I buy 2 audiobooks from Audible every month, and listen to them on my iPhone on my drive. Over the years, the books I enjoyed most are biographies, including the biographies of people (Tina Fey, Frederic Chopin, Louis Zamperini, etc), the biographies of cancer (Emperor of All Maladies) and economics (The Making of Modern Economics), etc. Having read the biographies of Alexander Hamilton and George Washington, it is hard for me to like Thomas Jefferson. However, having promised my husband to read the biography of Thomas Jefferson, I recently finished it to see his side of the story. Jefferson is a very complex and interesting character, with his wisdom and pragmatism, and his weaknesses (financial and slavery). What really intrigued me was how he recover from setbacks in life. As the war time governor of Virginia, Jefferson didn’t sufficiently prepare the state for the British invasion, and fled the state capital as the British arrived. His political career seemed finished in 1781. The next year, he lost his beloved wife of 10 years, and his life seemed finished. However, through the support of his family and friends, reading and science, horse riding and letter writing, change of environment, he gradually recovered his reputation and confidence, and achieve so much more later in life.

Being a scientist, it is natural that our scientific explorations experience ups and downs, and our careers ebb and flow. What I found to be very important is when things go well and opportunity strikes, we work like hell to move ahead. But even more important is that when things don’t go well and we are struggling to find a way out, we have the ability to lay low with grace. Just hang in there, make small steps forward, even stepping on the same spot or sliding a bit to shift some focus, but don’t drop the ball (掉链子) and totally loose grounds. What do I mean by “things not working well”? E.g. a course project got a bad score despite us working on it for over 30 hours, our teaching got a bad review, our experimental finding turned out to arise from an artifact, our paper got rejected or scooped, our grant didn’t get funded or renewed, an illness, a divorce, the loss of a loved one. What do I mean by “drop the ball”? E.g. too stressed out to even turn in the last project and going to the final exam thus earning a D in a course thus loosing PhD candidacy (turning in whatever you have done and taking the final exam with the best of your efforts could have earned a B- and save the degree), quitting science when one technology you are an expert on was overtaken by a newer technology (I considered quitting when finding myself working on the “dead technology” of ChIP-chip after my second maternity leave when next generation sequencing arrived), no show on progress meetings, lab meetings or in the lab completely when projects seem to be at a dead end, crying and blaming advisors for giving them the wrong project or not spending enough time on their project when paper got rejected, blaming lab members and collaborators under financial or publication pressure, disappearing from the lab for days or weeks, sabotaging lab members experiments or poisoning lab members (OMG!), etc. What sets people apart is not how much progress one makes in smooth times, but how resilient s/he is during setback. The ability to just hang in there, the ability to do the minimum at least to get by even when you are numb or hurt, the ability to lay low with grace, they will carry people in the long run.

What if you did drop the ball? Maybe we all did at different degrees occasionally, as long as it is not criminal offense. Don’t worry, time is a great healer. Didn’t Jefferson recovered from his governor fiasco? Just get back on your feet, and try again!

Feb 042015

Our Scientific Review Officer at NIH recently sent some notices about the Feb study section meeting, and the email includes a link to the CSR Peer Review Notes. I found the post on “Top 10 Things Reviewers Should not Say in a Review” to be quite interesting. I agree with most of the points, except the last one on “fishing expedition”. It is true “well-designed exploratory or discovery research can provide a wealth of knowledge”, but that should not be called a fishing expedition. When reviewers comment that some application is a fishing expedition, it usually means that the applicant just asks for money to do a lot of expensive genomics experiments without any careful design or idea what they might find.

The Peer Review Notes seem to have other useful information, so is a great resource…

Jan 122015

Many kids will learn some music intrument growing up and parents often need to invest time to help them with practice. I became a Suzuki mom after my older son Aaron started playing the violin two years ago. He is making so much better progress than me, and it is hard for me to keep up, especially now that my younger son Alvin started playing too. So recently I looked around for some tools that will help kids improve their practice efficiency, and found some happy surprises.

First, on YouTube, there are recordings of almost any piece the kids are playing, either from master recordings or student performances, like this masterpiece which I can’t wait for Aaron to start learning soon.

Then, on this website, if you paste in the Youtube link, it can convert the music track in the YouTube video into an mp3 that you could download into your own computer. I store the downloaded mp3 music in my DropBox folder, so I can access the file from my computer, iPhone or iPad.

Finally, a friend found this tool called AnyTune, which you can download on Mac, iPhone or iPad (it costs a little $). It can import mp3 music from DropBox, then play it out. Now here is the magic! It can play the music at any speed or pitch you want!! There are other functions like looping certain sections, placing location marks to go back to, but I found the speed dial to be the most useful. So kids can listen to the music (or accompaniment) at their speed, which helps them learn new pieces, understand the best pitch, tempo, articulation, and contrast. Now Aaron can put on earphones and play Vivaldi alongside Perlman at his speed, and his improvement is amazing!

With the professional music track in place, my motivation to practice the piano accompaniment for Aaron’s violin pieces has greatly decreased. So maybe I should practice hard to play the Bach double violin concerto with Aaron.