Predicting Causes of Drug Side Effects

Why do some drugs cause side effects?

A group of computational led by Peer Bork just published a paper in Molecular Systems Biology describing how to predict the causes of undesirable drug side effects by integrating drug-to-target and drug-to-side effect relationships and asking which protein targets are the most likely cause of the side effect:

For more than half of the investigated side effects, we can predict which proteins cause the side effects upon perturbation by a marketed drug. For the majority of these proteins, we also predict whether their activation or inhibition causes the side effect. 

The paper includes a pretty elegant proof of principle using a mouse experiment. 

The group predicted that increased sensisitivity to pain, one side effect of triptan based migraine medications, was likely caused by off-target activation of serotonin receptors, specifically HTR7.  They confirmed that treating mice with zolmitriptan (Sold as Zomig or Zomigon by AstraZeneca) increased their pain sensitivity and then showed the effect could be reversed by the SB-269970 serotonin receptor inhibitor.

The next level of prediction would be to know which individuals are sensitive to these side effects and which ones are not, probably based on the variations they carry in the side-effect targets.

Remembering to Forget

From time to time, we all set aside creative projects that can't be finished in one go.  It might be planning or doing something at work like drafting a paper or designing experimental strategies, to one of many personal projects like finding a neighborhood in which to live or simply deciding on a furniture layout.

There comes a time where one becomes fixed on details that aren't important, and the problem exists for it's own sake. 

When this happens, the project needs to be set aside for a day, a week, or a month so you can forget the unimportant stuff an return with a fresh mind.  In part, that's what a weekend is for. Every so often, details that were all encompassing on Friday yield to a solution on Monday morning. 

Remember that some things are best forgotten.

An Analysis of Almost 400 Public Private Partnerships in 2012

Brady Huggett provides a quick overview of partnerships established between academia and industry in 2012 at Bioentrepreneur:

Harvard University and the University of Texas system top the list of universities with the most industry deals, and University College London lands ahead of the University of Oxford and the University of Cambridge.

A few observations:

Until now, I wasn't aware that the University of Texas was such a big place of support for industrial partnerships.  No big Canadian collaborations made it into the tables.

Secondly, the type of collaborations are dominated by the cancer and infectious diseases fields.  Unfortunately, only 4% of the activity was in gene and cell therapy (My friends at Signals Blog will be disappointed!).

Finally, some of the bigger announcements from the past year are mentioned, namely the CureBeta Initiative translating betatrophin as a possible treatment for diabetes and the Fluidigm/Broad collaboration to establish a Single-Cell Genomics Center.

How Mature Was the Last Presentation You've Given?

Marc Kuchner, at Soapbox Science says:

We often argue over the quality of our colleagues’ presentations. When it is hiring time, for example, and faculty candidates are parading through your department, no doubt a common topic of conversation is who gave the best talk. And the maturity level of the research is often a contentious point. With these conversations in mind, I’d like to suggest a numerical scale we can use to describe scientific talks. This scale is not meant to weigh the overall quality of a talk, only to resolve some of the tension between those who prefer solid conclusions and those who enjoy more nebulous forecasting. 

The five point scale ranges from talks "that might one day lead to a testable new hypothesis or new data" to talks that "describe data and calculations that the community recognizes as part of its culture and history".

Though the simplicity of his five point scale is appealing, the higher levels depend on having a really great reputation and more importantly, time for that reputation to have developed.  So unfortunately, a Science Maturity Level 5 presentation is out of reach for most early career researchers.  However, it's still good to keep in mind that good talks don't have to be a flashbang hypothesis-data-conclusion-hypothesis-data-conclusion presentation that makes a scientist sound like an android from the 24th century.

Check out "What's Your Science Maturity Level".

From Gene Symbols To Financial Crises, Excel is There

Here's a thing that economists could have learned from computational biologists: Silly mistakes with Microsoft Excel can have a serious impact on your work.

In an op-ed at The New York Times, Paul Krugman shares the story of how a flawed Excel formula contributed to a paper being published by Carmen Reinhart and Kenneth Rogoff stating that debt-to-GDP ratios above 90% contribute to much slower economic growth and became ammunition for anyone ideologically bent on cutting government spending, for whatever reason.  Krugman, to his credit, is an author of popular books on economics like 'End This Depression Now!' and 'The Conscience of a Liberal', which you may have read if you follow him.

Krugman points out that the problem with the Reinhart-Rogoff paper didn't really surface until the financial research community attempted to replicate the results, but couldn't manage to do: 

Other researchers, using seemingly comparable data on debt and growth, couldn’t replicate the Reinhart-Rogoff results. They typically found some correlation between high debt and slow growth — but nothing that looked like a tipping point at 90 percent or, indeed, any particular level of debt.

Finally, Ms. Reinhart and Mr. Rogoff allowed researchers at the University of Massachusetts to look at their original spreadsheet — and the mystery of the irreproducible results was solved. First, they omitted some data; second, they used unusual and highly questionable statistical procedures; and finally, yes, they made an Excel coding error.

Whether or not their intentions to omit data or use unusual statitstical procedures is another subject, but it's clear that misuse of Excel was clearly found to be one of the culprits.

Which brings me back to biology.  Why does Excel change gene symbols to dates?

During my PhD, my group sometimes used Excel to format tables for presentations or reports, and we quickly found gene symbols that were automatically converted into dates or scientific notations.  The solution at that point was to fix them by manually escaping the field with an apostrophe as a prefix to that Excel would know to treat the cell as text.  So "DEC1" became "'DEC1" prior to importing the data.  Shortly after we realized that this was a prevalent problem in the field, we considered writing a short paper on it but deemed it too trivial for a paper.

Behold, a few months later, someone publishes a paper on the very same topic in BMC Bioinformatics.  "Mistaken Identifiers: Gene name errors can be introduced inadvertently when using Excel in bioinformatics" summarizes the problems nicely with Table 1 in the paper:

You can see that the Septins, the OCTs (Octamer-binding transcription factors), Deleted in Esophageal Cancer (DEC1 and DEC2) genes, and a few others are affected when converted to date format.

Symbol changes generally don't become a huge problem as long as the data isn't being exported and run through software that uses the gene symbols, but you can appreciate how manipulating biological spreadsheets in Excel can be a pain, especially if one isn't aware that escape characters can be used, or let alone what escape characters are.

Moral of the story: It's important to double and triple check your work at each step of the way, otherwise minor coding errors might affect important things like gene lists or government policies!

Two Linked Drugs Are Better Than One

The MIT Technology Review has a short article describing a new strategy for double dosing multiple drugs; linking them together:

Catabasis has found a way around the challenge of identifying a single active molecule that can hit multiple pathways effectively: using a chemical linker to bring together two active molecules. The synergistic effect of the linked molecules may arise from the fact that the two compounds both “get to the right place at the right time.”

The company, Catabasis Pharmaceuticals, is developing a technology that can combine two compounds together using a linker that supposedly puts the drugs into an inactive form in the bloodstream, but which is cut within cells when the whole package is absorbed.

CRASS: Create Relevant Abbreviations for Science Studies

Alex Bond at The Lab and Field, on the misuse of abbreviations in research:

Abbreviations obfuscate meaning, and create a separation between those who know what it means, and those who don’t.  It’s fine to use them colloquially, but for professional correspondence like a work email signature, or affiliation on a manuscript, it looks like crap, and won’t matter to anyone outside the abbreviated institution or country.  

I totally agree.  Abbreviations are thrown around from time to time, but really, if you're going to put out something permanent that people will refer back to, be it an email, a published paper, a conference presentation, you should spell everything out.  At very least once in the document.

How many times have you read a paper and found an undefined acronym?   Commonly used methods, like RACE or PCR¹, are especially vulnerable to the assumption that the reader knows what they are.  RACE, for instance, is an acronym with several meanings: "Rapid Amplification of cDNA Ends"; "Row-based ASCII Compatible Encoding"; "Return on Average Common Equity", and many more.

But acronyms (backronyms?) get even funnier when their force fit to projects:

And, for goodness sake, please avoid trying to make an existing lab/project/grant title into an acronym by selectively choosing letters to spell out a single word.  Nothing cries out “lame” so much as the Laboratory for Massive Experiments (LaME). You get the idea. 

I won't promise that I'll never make up an acronym that, but if I do, I'll try to keep it relevant.  When naming a scientific study with an acronym, remember to Create Relevant Abbreviations for Science Studies.

In the end, don't assume your reader knows what you're talking about.  You never know who will be reading what you put out there, so help them understand what you've written.


¹PCR, for anyone that didn't know, stands for Polymerase Chain Reaction.

Postdoc Income Taxes: How Much Should I Set Aside?

As a short follow up to my post taking you through a hypothetical Canadian postdoc's income tax return, a few colleagues have asked me how much should be set aside from each paycheque to have enough to pay the inevitable tax bill come April.

The strategy is extremely simple: If you're in a situation where your payor does not deduct income taxes from your pay, you need to do it yourself.  Transfer part of each pay into a separate account, like a high interest savings account, or send it to the government yourself (though most people will tell you to wait until April).

What I've done here is assumed that you're a postdoctoral fellow that receives T4A (Box 105) Fellowship income, has no other income, is single, has no dependents, and has not paid any income tax through the year.  These are the numbers my 2012 tax software (TurboTax) returns as the Balance Owing.  They should cover many people, but of course your own situation will be slightly different so I'd say this is best used as a  rough estimation of how much to save to reduce the pain at the end of the year.

I've also ran the numbers through TurboTax for three provinces: Ontario, British Columbia, and Quebec.

Postdoc Salary	Ontario Payable	BC Payable	Quebec Payable
$30,000	$4,117	$3,668	$3,110
$40,000	$6,292	$5,792	$4,363
$50,000	$9,332	$8,537	$6,012
$60,000	$12,447	$11,507	$7,849
$70,000	$15,577	$14,477	$9,686

Here's a quick example:

If you're in Ontario and are paid $40,000, you will pay $6,292 for the year.  If you are paid bi-weekly (26 pay periods per year) you need to set aside $6,292 / 26, or $242 per paycheque, to cover your income taxes.  If you're paid monthly, you'll need to part with $6,292 / 12, or $524 per paycheque.

I set up automatic transfers to move my estimated tax bill from each paycheque to an ING Direct account, so that the money is available when Revenue Canada wants it.  If you open an ING account using my  referral key (13876593S1), you can get a $25 bonus if you deposit at least $100.


Updated October 2, 2013: Included reference to ING Direct

Self-Underfunding Science

The Guardian has a nice comment on science underfunding being a cause of papers published lacking sufficient statistical power.  The point of more money leading to bigger and better studies isn't lost on many researchers, but underfunded studies create a much more serious problem than the occasional flaky paper:

The more serious problem with under-resourced neuroscience is that it systematically distorts the established body of supposed knowledge by increasing the number of spurious findings – so-called "false positives". Exactly the same argument applies to any empirical science.

But who exactly, is underfunding this research?  It's always easy to point the finger at a nameless, faceless, governmental funding body, and unfortunately that seems to be the status quo.  The other party involved in triggering an experiment is generally the principal investigator, who ought to have planned out the work in advance, received a grant for it, and set their group in motion.  The Guardian just manages to touch on this:

If there is a genuine commitment to funding a particular experiment, then it is essential that enough money is allocated for that experiment to be carried out properly.

If this isn't done in advance, there's a strong likelihood the work ends up cited, but as an example of how not to organize an experimental project.

Small Large Impact of The Proposed UK "Sequencing Tax"

Update: This post has been updated following conversations with @BioMickWatson.

North Americans are no strangers to the 2008 financial crisis and the austerity measures reverberating through government budgets since that time.

Over at opinionomics, a discussion has started on a proposed tax on UK sequencing services:

All of the UK academic sequencing facilities are in the same boat, and the consensus opinion appears to be that, after August 1st, we will need to start charging VAT (a 20% consumption tax) on sequencing for academic researchers outside of our own universities. The VAT of course goes straight back to the treasury.

This opinion stems from a short document very obviously entitled "VAT: Withdrawal of the exemption for business supplies of research between eligible bodies".  Here's the example cited:

A charity grant funds University A to carry out some research. The supply University A makes to the charity is outside the scope of VAT. However, University A needs to subcontract part of the research to University B. At present, the supply University B makes to University A is exempt from VAT but after the exemption is withdrawn it will be taxable (20% VAT). 

Sounds bad, doesn't it?  It's actually rather convoluted.  Read on.

The actual impact of the 'sequencing tax' will only affect transactions between universities, so while the example given by opinionomics is correct; academic sequencing facilities will start charging tax on contracts outside of their own institutions, some exemptions are unchanged. For instance, the UK consulation paper outlines the following relationships:

The withdrawal of the exemption for research supplied between eligible bodies will not affect the VAT treatment of research that is currently outside the scope of VAT (non-business research) or research that is standard-rated (business research).

So though it sounds like most not-for-profit grant funded research isn't affected by the VAT, that's actually not the case, according to @BioMickWatson.  He pointed out an even stranger situation created by two universities collaborating on a sequencing project: If grant funds flow to a university sequencing facility through another university, VAT is charged, but if the money is paid directly to the facility, it isn't. 

In Ontario, universities reclaim a large part of the HST they pay, dropping the net effect of tax from 13% to about 3.4%, according to the University of Western Ontario.  It seems that this doesn't happen in the UK.

And finally, in the case of UK sequencing facilities having to charge the tax, here's the most important paragraph to consider, which describes University B providing services for University A:

University A’s costs increase by the VAT charged by University B. However, since University B can now deduct input tax on making this supply, this may allow, in some cases, for a lower net price to be charged to University A.

So if the new sequencing VAT is anything like the HST for businesses in Canada, whatever has been paid on inputs used by the sequencing facilities should now be written off.

Here's how it works in Ontario: If a business purchases inputs costing $1,000 they pay $130 in consuption tax for a total of $1,130.  If they can't deduct input taxes, they will need to charge their clients at least $1,130 to break even. 

But if they can deduct the tax, they need to charge their clients at least $1,000 + 13% HST, or a total of $1,130.  The $130 tax paid on their inputs is deducted from the $130 they collect from clients, leaving nothing to send to the tax man, as they've already paid their suppliers the HST on their inputs. 

The net effect to the client will be zero, except for some additional paperwork, a small impact if you're in Canada.

In the UK, however, it seems the tax system is changing to one that treats research as a business without taking into account that in most academic cases, it's actually more like a charity or a public good.

Interactive and Open Peer Reviews Are Better For Science

Science Careers just published a two part article describing an interactive peer review system (part two is here), describing it as a chance to advertise one's work and network on one hand, while considering the novelty of the approach a drawback on the other:

Interactive peer review can offer advantages for researchers. Among the most important: turning a blind, mostly one-way process into a conversation with established researchers. "The primary benefit, particularly for young scholars, is getting their work not just out into circulation but out into active conversation with the people in the field," says Kathleen Fitzpatrick, a visiting professor at New York University in New York City who studies how networked communication technologies affect scholarship. But there are disadvantages, too: Some of these journals have no or low impact factors, for now at least, and more conservative scientists on review committees may not give you much credit for publishing there.

So basically what they're saying is that if you're going to consider publishing some work with a house like Frontiers, you might as well view it as an opportunity to advertise yourself rather than establish a toehold in a big journal.  That approach in itself isn't necessarily bad.

The two parts of the article really portray the collaborative review system as a powerful mechanism for constructive paper writing, which is a useful feature for people early in their careers.  However, there is one feature mentioned near the end that I think is move in a good direction; open peer review and the disclosure of reviewer names.

This is important on several parts.

Firstly, it puts much more power into the the author's hands, as it partly prevents reviewers from being overwhelmingly negative/sarcastic/arrogant/destructive which mostly ends up creating essentially useless reviews. They can't really hide behind anonymity, so are forced to at least try to be useful.

Secondly, and I don't think I've heard anyone discuss this before, is that it actually let's reviewers quantify their reviewing activities as professional contributions.  Most researchers have a section on their CV that includes which journals they've reviewed for, but it's frankly not very useful because a) I can't easily verify that you've been invited as a reviewer (I guess I could ask the editor), b) I have no idea whether you've reviewed more than one paper, ever, and c) I don't know if you're just rejecting everything that comes your way.  If everyone could see how much you're contributed to the scientific community, it might be an one more metric to evaluate productivity.

And lastly, if full peer review comments started to be published by many journals (some already do), it would be nice to read them to identify potential colleagues that are great partners to work with and also the people you would rather avoid.

Sequencing Companies: Infrastructure of The Next Medical System

It sometimes pays off to go through your notes, however old they may be.

After AGBT 2013 and the associated flurry of interest in clinical exome sequencing finally making it, I did a bit of research into what companies would become 'infrastructural' in this new medical reality.  Somehow, the notes were buried under other work and resurfaced after I went on a desktop archaeological dig.  It was bad enough to make David Allan cry.

Out of all of the companies I noted that Illumina, in a relatively quiet way, issued a press release on it's acquisition of Verinata, developer of the verifi prenatal test which tests maternal blood for evidence of a few major chromosomal abnormalities, like Trisomy 21 (Down syndrome), Trisomy 18 (Edwards syndrome), and Trisomy 13 (Patau syndrome).

Illumina announces agreement to acquire Verinata Health, a leader in non-invasive prenatal testing bit.ly/Wn9wbf
— Illumina (@illumina) January 7, 2013

Curiously enough, news of Verinata's acquisition was made public in January and was completed during AGBT, yet no mention of Verinata (that I noticed) at the conference was made, despite it being a large purchase ($350 million + $100 million in milestone payments), and despite the obviousness the conference projected of clinical sequencing arriving in full force.  The conference wasn't even finished and speculation that clinical reports are going to be the theme next time around started circulating on a newly minted #AGBT14 hashtag.

Consider the expanding role of clinical sequencing with the dominance of any of the sequencing providers and you can quickly see that there's an opportunity to position themselves as infrastructure in a clinical sequencing market.  It's a nice position to be in.

When it comes to television, Comcast sells you the cable service and lets networks compete to sell you programming.

When it comes to clinical sequencing, will Illumina rest happy doing the sequencing and let biomarker providers battle it out in the market?

Irreproducible Research Isn't Always Useless

In Nature, Jonathan Russell argues that something needs to be done about reproducibility of papers:

In recent years, it has become clear that biomedical science is plagued by findings that cannot be reproduced. This wastes grant money and hinders development of new treatments and cures. 

Yes, but the true answer depends on why the study irreproducible.  Is it because the authors have been dishonest with their data or is it because they were truly trying to publish something that's completely novel?

If a paper cannot be replicated, the authors should be required to amend or retract it. Funding agencies would then consider a principal investigator’s history of reproducibility in grant reviews. 

Funding agencies should already consider whether the author has had too many errata published, and on what basis each of them came to be.  There are many examples of errata published for honest mistakes like mislabeling figures, author name corrections, etc.  Fundamentally, this level of added scrutiny has to come from the funding agencies, perhaps by providing better incentives for qualified reviewers to, well, review.

I don't necessarily agree that these papers should be outright retracted (unless the paper is fraudulent), because there's always some good message to take home, even if it something like "don't ever use these experimental conditions because you will end up like this team of researchers".

If publishing amendments were explicitly rewarded, people would take the time to do it.

But the point of doing good research is to get it right the first time.  The explicit reward already comes in the form of publishing something that confirms or refutes the first study.  Plus, the team that publishes the 'amendment' (really, just another paper) get credit for doing so, in the eyes of granting agencies.

On principle, I agree with Russell that improvements are needed to identify irreproducible work, but more on a level of catching dishonest folks (which Retraction Watch sort of does, after the fact) instead of bulking up the scientific system with more administrative controls.

If academia were another type of business it would be...

A few days ago, I was having lunch with another scientist and we had a friendly debate about what kind of business was most similar to academic research. In the end, we couldn't settle on a single sector, but we couldn't agree on whether 'academic research' was it's own standalone style of business -- there are too many common themes with other parts of the economy.

Though there are many industries which can supply strategies useful in academic research, none are a perfect model for academic researchers.  But many do offer concepts that can be borrowed:

Advertising.  Part of doing academic research is getting your message out there and using papers, conferences, and media to promote your work and your lab.  Like it or not, your lab is your brand.

Biotechnology/High Tech.  If you're in life sciences, a biotech model pushes you to be efficient to produce that final widget that gets you to your next slice of funding, whether it's another investment or a grant.  Similar analogies can be made for other faculties like physics, engineering, or chemistry.

Education.  Teaching models exist that can help train trainees and other members in your research group.

Legal.  Positioning your research as an advance in the context of literature is a little like putting together a convincing legal argument.

Service Industries.  If you've ever collaborated with another research group it's useful to treat them like a valued client and deliver on what you promised, quickly, efficiently, and to at least the standards expected.

Antibody-drug Combinations Delivering in Many Ways

Here's a nice review of the state of antibody-drug combinations (ADCs) as anticancer drugs, written by a team at Trinity Partners. Published in Nature Reviews Drug Discovery today, here's the highlight:

Genentech's ADC ado-trastuzumab emtansine was approved by the FDA in February 2013 following the positive outcome of the Phase III EMILIA trial. In this trial, patients with HER2-positive metastatic breast cancer had an improved median overall survival benefit compared to the combination of lapatinib and capecitabine. Genentech licensed ImmunoGen's Targeted Antibody Payload ADC technology — which includes their proprietary linker platform attached to DM1 — and combined it with the blockbuster HER2-specific monoclonal antibody trastuzumab. DM1 is a highly potent cytotoxic that is 100 to 10,000-times more potent than standard chemotherapies. As ADC technology enables selective targeting to tumour cells, ado-trastuzumab emtansine is highly effective and has less toxicity when used in this formulation compared to unconjugated DM1.

For what it's worth, DM1 is derived from other compounds that inhibit microtubule formation and thus cell division and is in the family of compounds like paclitaxel.  It's an agent that generally kills any cell it approaches, and based on it's potency I don't think it would make a useful chemotherapeutic on it's own.

Nevertheless, what makes the concept of antibody-drug combinations very appealing is the modularity of the approach:  You can focus on development of either the drug component or the targeting antibody.  Remember, the drug being delivered need not be the atomic bomb that is chemotherapy; and an antibody-drug payload technology still leaves space for development of disease specific drugs.  ADCs simply put them in the right place, and make for a nice neat ecosystem for researchers to operate within.

The Canadian Postdoc's Guide to Income Tax Returns

Doing income taxes aren't really a subject that excites most people, much less so if you're a scientist.  But ask any postdoctoral fellow or PhD student about income taxes, and you'll usually find someone very excited to tell you why they should or should not be paying taxes, depending on whether they consider themselves an employee or a student.

The debate about academic researchers and taxes in Canada was a hot topic for several years starting in 2006, when a scholarship exemption was introduced making training stipends tax-free.  Almost instantly, you could hear PhD students across the country breathing a sigh of relief once the onerous obligation of paying about $1000 a year in taxes on their T4A Scholarship & Fellowships income was erased.

Postdoctoral fellows were similarly happy because they too, became exempt from taxes on their relatively larger salaries (sorry, stipends).  All they needed was a T2202A form to certify they were in an educational program, which many universities were happy to provide.

But the University of Toronto was one of the cautious few, refusing to issue those precious T2202A forms to postdocs between 2006-2009, to the displeasure of many and triggering a lot of debate across Canada.  In an unfairness that rankled many trainees, some postdocs paid taxes and some did not.

In retrospect, UofT's move looks like it was a wise one.

In March of 2010, the government effectively shut down this manna raining down on trainees with it's simple budget decree that "post-doctoral fellowships will be taxable".  Later that year, the Canadian Revenue Agency also stated that they would not accept that PDFs were qualified for tax exemptions between 2006-2009, which raises the possibility of many ex-postdocs becoming liable for back-taxes, some to the tune of 20 or 30 thousand dollars. But so far, I've never heard of anyone having to pay.  Thankfully PhD students get to enjoy their tax-free stipends, for the time being.  My friend David Kent has also put up an article summarizing the history of problems with postdoc taxes over at the Black Hole.

Anyways, I'm digressing; I'm in the camp that once you're out of an educational program, you're not a trainee, you're an employee.  While it's true that a postdoctoral fellow is given opportunities to learn technical skills, develop as a professional, and learn to manage their own projects, it's fallacy that a postdoc is different than many other middle-level career positions where you get to experience the same or very similar kind of personal growth.  I have many friends that went through medical residencies, articling, or obtained their CMA or CA designation while working in various companies.

All in all, being a postdoc is a pretty decent job, but there are many catches in the tax code that are annoying if you don't think financially, and so whether you're a postdoc or student, it definitely helps to educate yourself on financial basics with elementary books on the subject:

• Graduation Debt: How to Manage Student Loans and Live Your Life for advice on paying down your student loans, 

• The Wealthy Barber Returns, a remake of the classic Wealthy Barber, which I read back in early high school and set a pretty good foundation for my whole post-secondary career (this one is out of print and hard to get), or

• The Millionaire Mind, one of the best books I've read on understanding the psyche behind people that are financially independent.  A great set of results from studies conducted by Thomas Stanley, PhD.

If you're a postdoc (which you probably are if you're reading this), I'd say that general financial planning books advising on buying a home or making investments are probably going to be more useful over the usual ones encouraging you to get out of debt by cutting out lattes, or using a system of envelopes to keep your budget.  You made it through grad school and I'm sure you already know how to cut expenses to the bone.

Anyhow, let's run through some taxes.

My own tax return is a little more complicated than I describe here (I do some consulting from time to time), but I'll work through my tax return as a "typical Canadian postdoc" residing in Ontario, and earning only fellowship income in 2012:

Line 101 - Employment Income (Box 14 on T4 slips)
Despite working like a dog throughout the year, I didn't get any employment income!  This income is considered "earned income", and becomes very important throughout the return.
Zero.

Line 130 - Other Income
Here's where you put your salary (sorry, stipend!) amount in from Box 105 on your T4A.  You get a basic tax exemption on $500 of that stipend, with the rest of your postdoc salary being taxable.  The "Tax exemption for full-time education" will remain Zero (unless you receive a T2202 form, which you shouldn't get).  This line also curiously leaves you with the impression that the government doesn't really deem a postdoc an educational activity.

Line 206 - Pension Adjustment
Pension!  Ha!  Usually this line contains an amount that adjust your RRSP contribution room downwards to compensate for money your employer paid into a pension plan for you.  Postdocs, move along now.

Line 208 - RRSP Deductions
Ok, so I have no pension plan.  Perhaps I can contribute to an RRSP, reducing my tax payable and take the responsibility of saving for my retirement.  Here, sadly, you need to have "earned income" to generate contribution room and a salary on a T4A doesn't count.

Line 300 - Basic Personal Amount
The first $10,822 that everyone makes is tax free.

Line 367 - "Children"
I have one child, who brings me much joy, much more so than reading the latest papers in Nature or even presenting my data at a conference.

And because I have that one child, I get to deduct another $2,191, which represents a whopping tax savings of about $330 per year.  (I can assure you that my son eats much more food than this will ever buy.)

Line 308 - Canada Pension Plan Contributions
For postdocs: Zero.  Employees earning up to about $50,000 per year will max this out at $2,306, and if they don't feel like they would pay enough they can complete Schedule 8 to voluntarily contribute to the CPP (The logic of doing this is a matter for other websites to debate).

Unfortunately, like with the RRSP example, a postdoc's T4A Fellowship income doesn't count as pensionable income, so you can't even make use of this.

This is particularly annoying as CPP payments are calculated from your best 40 years of earnings between 18 and 65 and many postdocs in the 30-35 range have essentially "earned" nothing throughout undergraduate and graduate degrees and forfeit part of this government benefit forever.  Arguably, advanced degrees should make up for this financial hit, but there are no guarantees.

Line 312 - Employment Insurance Premiums
Here we have a similar situation to CPP, above.  Postdocs need not apply.

Why bother with Employment Insurance, you ask?  A research position isn't a strenuous job: Yes, there are some significant health hazards, but postdocs aren't lifting heavy objects, are probably not going to fall off a roof, and generally don't have many workplace injuries, except for pipette thumb and carpal tunnel syndrome.

But the one major benefit of EI is parental leave, and this is one benefit that seems important to the postdoc (i.e. 26-35 year old) demographic.

While it's true that some postdocs are entitled to parental leave, the duration is either four or six months depending on whether the parental leave comes from NSERC or CIHR funds, respectively.  The University of British Columbia further highlights the challenges facing postdoc parents-to-be, as there are different rules for whether one is a 'trainee' postdoc or 'employee' postdoc.  The latter pay into EI and are eligible for the full 12 months of combined maternal/parental leave, which pays an amount equal to 55% of your salary.

Progressive policies notwithstanding, I haven't heard of any postdocs going on parental leave using NSERC or CIHR funds, but I would like to hear from anyone that has actually gone through this process.

Line 363 - Canada Employment Amount
This would have been $1095, but since a fellowship isn't classified as 'earned income' we'll leave this at zero.
Forgone tax savings: $164.

Line 364 - Public Transit Amount
Do you use a monthly transit pass to get to work?  Write it off!  You'll get 15% of their value back in tax reduction.

Line 319 - Interest Paid on Student Loans
Are you still paying off your student loans?  There's no shame in taking your time to get that loan balance down to zero, but in the meanwhile remember to claim the interest you pay here.
Value: Up to 20% of your interest paid.

Line 323 - Your Tuition and Education Credit Amount
For early postdocs, you've probably accumulated a lot of unused Tuition and Education amounts from your seven five year PhD.  Under the current tax rules, a PhD student gets a credit of $465 per month, plus the value of tuition paid, totaling about $12,000 per year as a credit used here to reduce tax owing in the current year.

Since the 2006 scholarship exemption kicked in, students just carried these credits (worth about $2,400 in tax, in Ontario) forward for use in the future, when they actually don't tax-free stipends or salaries (like postdocs).  This will usually cover your first year of postdoc taxes.

We're nearing the end!

Line 435 - Total Tax Payable
This line gets filled in after you've completed Schedule 1 (for Federal Tax) and Form 428 (For Ontario Tax), and basically indicates what you should have paid in tax throughout the year.

For a single individual in Ontario with $40,000 of T4A Fellowship income it's about $6,300.  With $50,000 your tax bill is about $9,300.  And if you're making bank with a Banting Fellowhip, collecting $70,000 from the government, you now get to send $15,600 of that back!

Line 437 - Total Income Tax Deducted
Finally, this is probably the most terrifying line for many postdocs, because it's probably Zero, as in "You've paid Zero to offset the sting of Line 435".

To my knowledge, you can't elect to have income tax deducted from a T4A pay stub (I've tried), and while it's true that you can opt in to paying installments throughout the year yourself (Line 476), I've only met one postdoc where the CRA has sent them a letter requesting that they do. Instead, most tax-aware postdocs have an automatic savings plan through a bank like ING Direct to set aside funds for the tax bill the following April.

Unfortunately, many tax-unaware postdocs won't have done this and experience the nausea of realizing that they need to find several thousand dollars to remit their back-taxes before April 30th.

So don't become one of those postdocs.  I've put together a table of total income tax payable for postdocs in Ontario, BC, and Quebec that serves as a starting guide for how much you might have to pay.  Save it throughout the year. 

I set up automatic transfers to move my estimated tax bill from each paycheque to an ING Direct account, so that the money is available when Revenue Canada wants it.  If you open an ING account using my referral key below, you can get a $25 bonus if you deposit at least $100.

That pretty much covers the basics.  If you have anything more than a T4A and rent payments (which I didn't include above), I do recommend using tax software like TurboTax for the time savings over using pen and paper.  This is especially true if you have a spouse, have some savings and/or investments, collect rent, etc.  I also need to stress that this article isn't tax advice and if you're in doubt of what applies to your own situation you should consult your own tax advisor.

Good luck!

TurboTax:   This is what I use every year and it's highly recommended.

Other personal finance books you might find helpful include:

If you're an American contemplating a move to Canada:

Update July 30, 2013: Included a few words on personal finance.
Update October 2, 2013: Include reference to ING Direct

"Because That's Where The Money Is"

Here are some interesting facts:

The top 22 life sciences Tool and Technologies companies spent over $6 billion on research and development in 2012.

By comparison, the annual NIH budget is about $30.9 billion.  The CIHR budget (for a country 1/10th the size of the USA) is about $1 billion.

Don't Be a Hero

Christopher Fasano describes how his lab is great at growing neural stem cells but not very good at genomics and bioinformatics:

I can surely learn this process and try to integrate it into my lab. But that would be heroic, and as you know, I am not a hero. So what do we do? We collaborate with bioinformaticists and mathematicians who do this for a living, and we all come together to move the project along.

Everyone starts out having that urge to "be a hero" and save the day, however small.  Some people learn to control it early and some people go on trying to be a hero in various circumstances for years. 

Doing a few years of research teaches people to be more collaborative as a natural by-product of dealing with the unknown.  It has a relentless ability to bring down people trying to throw wild pitches and make complicated experiments work. 

Of course, if the results are great the student/postdoc/scientist looks brilliant and becomes the hero of the day.  But remember that these heroic stories come at the cost of hundreds of failures that bulk up a dusty lab notebook somewhere.