Learning academic writing

Each year at The Institute - ir's coming up to five years in January - I've been supervising final year research projects: sometimes a handful, sometimes a dozen, one year 16! The kids have a choice of doing this task in the lab "all wet" or with me using computers "dry" to reveal the pattern and process of evolution. The 'Binfos' are a mixed bunch - some, like me, known to be inept at the lab bench; some slackers looking for an untaxing option; and every year tuthree who are really the best students we have - curious, motivated, self-starting and independent. I love 'em all! Not least because of the mix - getting our least academic students to fulfill their potential and a little bit more is just as rewarding as having an adult discussion about the evolution of 'flu viruses or the epidemiology of  Huntington's Disease. Whatever the level, these are all original research projects - finding out something about the world that nobody else on the planet knows. That's a pretty rewarding challenge for someone who has just got the vote.  How do we know that nobody has been there before? Each student is requested and required to carry out a review of the scientific literature on some part of the natural world. Sometimes, especially with our rocket scientists, I will be bullied into 'supervising' a project which is driving the student. Otherwise, I'm allowed to let my butterfly mind flit over the meadows of knowledge looking for a pretty flower to investigate. It must also be admitted that I'm not above whoring out my Effectives to answer questions posed by colleagues and collaborators. That can work out really well, and at least one student has parlayed that relationship into a Master's degree elsewhere.

It's nearly the end of term and by 1700hrs Friday 8th, all our students, wet and dry, had to submit a first draft of their Lit Review. I then had a really interesting half-week's work reading, critiquing and returning them. This year the reviews were all serious pieces of scientific research, neatly capturing a variety of different topics. Even the essays submitted with a "this is crap, but I'm drawing a line under it here" warning were on-message, coherent, unwaffly and full of interesting stuff. So I've learned a lot, and we should all be collectively proud. In a number of cases, the LR has thrown up potential avenues for research that I had not thought of.

I won't pretend that all the work was written in perfect grammatical English: we have two Polirish this year and they are quite mean with their definite and indefinite articles. And don't get me started on the apostrophes scattered like confetti at a wedding. But this is really the first substantive [5000 words] piece of academic writing they've ever submitted. Accordingly I've had to explain some of the basics of communication.

They read, they write, I read, I come across a strange word. Now I had an expensive education, some of it involving Latin, so I can often make sense of long words. But sometimes I'm totally foxed so I Google it up and find a definition there. That's not really good enough. The Lit Review should aspire to being self contained, so I push the students to include a Glossary.

Glossary - should include all the words that only appear in your report, which you wouldn't expect your bff or another Binfo to know; because you're the only person on Campus who has read what you've read. In some cases, that will be an essential addition for the readability and utility of the report. Remember that the report goes off to an external examiner who is a generalist. S/he may not be able to spell Blast, let alone Huanglongbing or Sporolactobacillus. The glossary also serves as a memory aid: that three letter acronym TLA which you defined some place earlier in the report; I've forgotten what it means. All those should be in the Glossary.

They read, they write, I read, I come across a strange idea which I'd like to follow up. If it's come from the literature, there should be a citation embedded in the text which will kick forward to a list of references.

References.- Everyone should use referencing software like Endnote Mendeley or Zotero to match the citations to the references because it is part of the training. But the result has to be fit-for-purpose which is to allow the reader to follow up one of the statements to the source. In the old days of print this required: Authors, initials, (year), Title, Journal, Vol, pages. Back in 1977 I could write Smith AB 1975 NAR 213: 410-417 in my notebook and hunt out the relevant volume in the library. The students of The Institute are submitting a hard copy, so all this info shd be included for that reason. In the modern world of epub and full text, for any biomedical paper, it is handy to include the Pubmed ID like PMID: 8441625 because I can easily get to the source then. Referencing software operates on a GiGo [garbage in garbage out] principal and some of the kids have driven the software so that itmangles the author name or the citation, or doesn't include page numbers - it is part of the learning experience to sort this out.

As our reading gets so cluttered with hypertext links, I like the idea of being forced to think that a written essay is a mode of communication. To reflect on what must be done to ensure than your reader is not misled and is empowered to follow your thoughts forward into the unknown.

Top Ten Human Genes

Peter Kerpedjiev had an idea that required some moderately high-throughput analysis and got himself a full-page spread in Nature, Europe's premier general science magazine about The most popular genes in the human genome. This requires a mash-up of two sorts of data which are effectively orthogonal to each other - related but not correlated. Most popular is here defined by those genes which have appeared most often in the recent scientific literature. There's all sorts of other stuff we know about genes and their protein products: molecular weight; genomic location; which tissues they are expressed in; whether they are receptors or enzymes or signalling molecules or proteins that switch on other genes. You wouldn't expect that any of these things-we-know would tell us about the other attributes.

We had a paper in the 00s, for example, which showed that genes expressed in the liver (or heart or kidney) are scattered all over the genome. The long genes aren't all found on the longest chromosome.  Olfactory receptors are clustered in little groups, it is true, but there are lots of these OR clusters and toll-like receptors TLRs are all over the shop. We'd be mad if we only wrote papers about receptors and ignored enzymes. Kerpedjiev was curious about which genes/proteins occupied the collective time and energy of science and wrote a simple-enough script to snag this information for each and every one of the 27,000 protein coding genes we know about. It's exactly the same idea as I've been progressing with the Masters of Imm up in Trinity over the last several years from 2012 until they sacked me in 2016. I called it the Most Sexy Immuno-protein competition. We didn't aspire to be comprehensive because we couldn't write a simple-enough script without a lot of help. Nevertheless, we showed that some TLRs were stupidly more popular than others because science puts a lot of handicaps on doing original research: everyone - HoDs, funders, editors, reviewers and referees - is happier if you mullock along in the footsteps of others. Wenceslas science, we might call sing it.

A few proteins acquire legs and outstrip their trudging  rivals for the attention of scientists. Aled Edwards from Toronto did a similar study ten years ago in which he showed that the $1billion Human Genome Project had been effectively useless in generating new targets of research to ameliorate the human condition. Researchers found it easier to fondle each other's work than to strike out into the unknown. Working within the herd is safe but not very exciting. Going all maverick makes funders nervous and the results require too much effort to assimilate and tend to get ignored. I could ask you to guess which genes are most highly cited in the scientific literature, but even if you are full time in bio-science you likely won't have the breadth of interest to know them all, let alone put them in the correct order.

Well here they are [L]. I'm surprised that TLR4 isn't there but that's only because we discovered a minority interest TLR and so I think that TLRs are bound to be interesting to everybody and I acknowledge that TLR4 trumps our 'umble TLR15. But all TLRs are collectively a bit of a side-show. Even among the top 10, p53 is Eclipse first, the rest nowhere, but in a way that is reminiscent of Zipf's distribution laws for letters or Benford's for numbers. So who are these celebrity boys and girls of biomedical world?
p53 is the guardian of the genome a tumour suppressor which is found to be mutated in about 50% of all cancers. The implication is that, when fully fighting fit it is preventing the development of cancer.. Several of the other genes reflect the biomedical world's obsession with cancer - which affects the family and friends of the affluent white males in power in the West - rather than possible targets for infectious diarrhoea, TB or malaria. The million black babies a year who  succumb to each of those diseases can't afford to pay for drugs.  #2 is TNF whose name tumour necrosis factor says it all: it works to gee up the immune system to kill tumours. VEGF vascular endothelial growth factor is the source of another nifty insight to treating cancers. As a tumour grows through out-of-control cell division it demands to have more oxygen and glucose to fuel its energy demands. If we can suppress the develop the growth and development of the local matrix of capillaries then we can suffocate the traitor in our midst. EGFR epidermal growth factor receptor works on the same process from a different angle. If we can jigger the receptor of a growth factor, then we can also suppress growth. Note the GF in TGFB it's another target for growth factor control. Note the R in ESR1 the oestrogen receptor which is involved in ovarian and breast cancer and tell us the oestrogen has more role in life than just shedding an egg a month for 30 years.

I'll refer you to the original article for a really neat graph of the timeline of trendiness. When I was in college 3% of all the publications were obsessing with beta haemoglobin HBB, mutations in which led to the first genetic disease sickle cell anaemia. Hey, before President Richard Nixon signed the National Cancer Act in 1971 and diverted $billion$$$ towards the War of Cancer, biomedical science cared about black babies. You have to talk % impact because in absolute terms publications were a trickle back then compared to the tsunami of tosh today. I say tosh because the average citations for a scientific paper is less than one: more than half of all papers published bob up and promptly sink without trace effectively unread by everyone including the authors.
1980 280,000
1990 410,000
2000 532,000
2010 940,000
2016 1,259,000. Heck and jiminy there are even 13,000 pubs for 2018 out there - I guess mostly from the Journal of Clairvoyant Studies and Prognostics.

Experts

I love watching or listening to experts. People who pwn a part of the knowable universe. Anything will do: people who collect match-books; people who know all about the Taft family; people who know where fish will rise; the man who knows all 6,000 genes of the yeast genome; a child who has trained her dog.

• Josh Waitzkin was chess prodigy who, aged 11, drew with Garry Kasparov Га́рри Ки́мович Каспа́ров in a tournament, won the Junior Nationals and then stopped plaing, started doing Tai Chi and wrote a book called the Art of Leaning. here's the Executive Summary.  The pursuit of innocence.
• Just occasionally in Rugger one player sees the gap and runs a long way to the goal-line with a defender in hot pursuit. It's called a foot-race.
• I've played Scrabble since I was a nipper but I've never cleared my rack. Scrabble champs think differently.
• A very English furniture-maker.
• Patchwork, thrifty
• Scything the grass
• Splitting wood with either hand or with both hands (and added grunts)
• Those with a college degree are not experts - they know nothing. But because they have lasted the course they land the desk jobs that pay well. As Dau.II, who has spend the 1st four years of her voting life working in U.Life, says "No degree is the new degree".
• Knowing one book well might make you an expert. You can Call up Ishmael and tell him about it. You'll be youtube-famous.

Birthday lies

It's The Boy's birthday today. In contrast to the arrival of his sisters I wasn't there, then. My maternal grandmother was christened Lily Valentine because she was born on 14th February 1893. Even back then Valentine's Day was a thing. I was glad she had that birthday because I still remember it and she's dead nearly 17 years. It's like being born on Christmas Day and getting called Noel. Actually, as I pointed out in 2013, not many people do get born on Christmas Day, especially in 'advanced' countries like the USA because of elective C-section. That allows you to pick your delivery date so it doesn't interfere with the school holidays or a vital business meeting at work. But also more positively, there is an up-tick of births on Valentine's and Paddy's Days for strictly romantic reasons. Follow the link to see if you were born on a minority bday. Later on I did analysis that showed quite clearly that births are not distributed evenly through the year but experience a boost in August and September. It can't all be because people get drunk over The Holidays 40 weeks earlier. It's more likely an evolutionary relic of coming into heat at a time the ensures birth at the best time for survival: mother and baby doing well. Sheep have a much more fluctuation in birthdays.

Years ago I signed up for a feed from www.quora.com and read the traffic there far more than I should. Quora is a stream of inane questions, interspersed with kids expecting adult strangers to do their homework for them but it is 'mostly harmless'. Occasionally something interesting pops up:

That was from Jérôme Cukier's answer to the question "What are some surprising or non trivial statistical facts about Facebook users". I'm one of the last holdouts against the Empire of the Zuckerclones but I thought one of the terms of service was that a) you have to use your own name and b) you had to be over 18. That's fatuous because unenforceable. Facebook also asks for your birthday, ostensibly so that their robots can send you an e-cake with pixie candles on the day and make you think that somebody cares. Well it seems that, in the rush to get registered, people put down any old date when asked with notably preference for the 1st of any month or 01/01 02/02 etc. dates.  Particular /peculiar chosen dates include 4th of July and Valentine's Day. You can maybe see the shadow of the [real] September births are up trend.  You should always take the results from questionnaires with a skeptical pinch of salt.

Designers of census forms please note. One of the anomalies thrown up was the existence of a number of teenagers who had endured multiple pregnancies. The most likely explanation was that ages like 76 and 79 had been tallied up as 16 and 19. Most of the time it doesn't matter much because that's what standard error is for. All those blips up and cancelled out by blips down and the Census finishes up correct with an allowable +/- error of estimate. Sometimes the errors are from laziness and ineptitude of the enumerators or the punters. Then again, sometimes it's because some people are a little hazy about the facts of life. Recent surveys have found 5 women in every 1000 who have been pregnant but have never had sex. The same source reported a few virgin fathers but that smacks of cuckoldry.  Similar transcriptional errors recorded in The Blob

Say Cheese

"Limburger"
You can spend a lot of money on cheese or more modest amounts. Experts can tell the difference. Whichever, whatever, I love the stuff, since I was first exposed to different cheeses aged 8. I've had a good bit to say about cheese on The Blob

• Why Limburger smells of feet - it's the Brevibacterium linens silly.
• Grandfather's pumpernickle, red onion and Limburger sandwich.
• Modern commercial variety
• Kids give it a go: "Did someone fart in your mouth?"
• Cheese rind is well safe. But not the wax on Gouda, y'dork!
• Cabrales is wrapped in leaves.
• How to make Jarlsberg.
• Silke Cropp makes Corleggy cheese in Co Cavan. When we lived in Dublin 20+ years ago, Silke would bring her amazing cheeses to the Dublin Food Co-op and I'd buy a slab, or even a whole kilo at £8.50, every time I was there.
• Mini-documentary on Alpine cheesemaking and dairying.

While it may be safe to eat the rind but it's not always safe to eat cheese. The main problem is Listeria monocytogenes. This Gram-positive bacteria grows well in the presence or absence of air. It is common enough in the soil, and so is spread by plants that have contact with the soil. We're still pointing at canteloupes because a 2011 outbreak of listerosis was associated with that fruit [map R]. That deadly incident dealt a serious blow to canteloupe farmers in Colorado and elsewhere in the USA, but it was a once-off as far as Listeria and canteloupes are concerned. It's easy to wash the fruit - ammonia or ethanol will do it - because it is big and nobody eats the rind so you'd be ++ unlucky to get sick from that source.

Another once-off food contamination scare happened in the run up to Christmas 2014 with toffee apples [L for cases map]. The problem with apples is that the Listeria can lodge in the holes at each end. But that's okay because most people [self is excepted because I eat apples pips and allll] leave the 'core'. The toffee-apple case was unfortunate because driving the stick up the apple's oompah transferred the Listeria to the anaerobic interior of the fruit, where it grew anaerobically.  But most of the outbreaks of listeriosis recorded by the CDC in the last ten years have been caused by cheeses of various sorts.  A useful graphic for the process of determining  which puka-puka people are part of the same event.
Listeriosis could be a case study from Risk Assessment. This is the idea where you have balance the likelihood of an event and its severity to decide what sorts of events are worth preparing for or acting to avoid. Risk is usually defined as the product of the likelihood and the severity. Salmonella and Campylobacter contamination of food is common but 'case fatality rate' is only 1% while Listeria is much rarer but 25% of contractees seem to die. This is why health professionals talk large about Listeria. But the talk is always in the context of small babies (and late stage pregnant mothers; the elderly and those who are immuno-suppressed including those with AIDS and those who've had a hearth, liver or kidney transplant. That's because the rest of us has an immune system which can duff up Listeria before it gets started and that's why Listeriosis is rare.

Listeria is also a problem because of the wide range of temperature at which is grows. It is happy to go forth and multiply at 37^oC but will also grow and divide at 4^oC. That's what your fridge is set to! Cooling food and left-overs is a really effective strategy for slowing the growth of E. coli, Salmonella and Campylobacter. But cold-loving  'psychrophiles' like Listeria and Pseudomonas won't be bothered by the chill. It would be stupid to throw up your hands and say "Why bother with a fridge?" because that controls the growth of the vast majority of food spoilage microbes.

From pasture to plate

Mapping the intestine

I grew up in the Genetics Department in TCD where we were taught to 'treasure your exceptions' - that variety, difference, was where the interest was.  The rival department was Biochemistry and I characterised their whole endeavour as "They go to the abattoir and get a skip-load of calf thymus [see Right], run it through blenders and sieves, purity columns and filters to get 2 grams of active principle. They then write the paper as discovery of the mammalian enzyme for wuggawugganess". They ignored all the lovely, intriguing, interesting variability!! Not only any difference among the contRIPutory calves, but also among the different mammalian species which last had a common ancestor 60 million years ago.  Nevertheless, it was a similar bucket scale protocol that isolated insulin as a treatment for diabetes; so it wasn't a total busted flush as science goes.

Fast forward 20 years and I was employed in the Biochemistry Department, pushing different frontiers of science and realising that I'd have to eat my earlier sniffiness because my new colleagues were every bit as sharp as the lads from Genetics. The Gaffer was the Professor of Comparative Immunology and I brought an evolutionary / genetics view-point to the project. Which was to investigate how different species had different immune responses to the same insult, probably because each lived in a different ecological niche and was beset by a different cast of microbial thousands. OR to investigate how the same species reacted differently to different pathogens. OR both . . . although was going to be complicated to set up. In about 2007 our second grant application to the Department of Agriculture was for a comparative immunology study to measure the immune responses of chickens to a virus, a bacterium and a protozoan parasite. That would have addressed some of the subtlety of immune response. But as we costed out the trials, the replicates, the hands to run the equipment, their taxes and social security contributions the numb€r$ far far over-topped the limits set by the DoAg and we had to pull in our horns. In the end, 10 person years of solid work and €500,000 of tax-payers money made some modest contribution to understanding what happens when chicks are exposed to the bacterium Campylobacter jejuni. With the limited funding and clunky tools-of-the-day, we hauled out the caecum from a bunch of chickens - some infected and some not - and analysed its contents (count them bugs) and also measured which genes in the caecum were being switched on (and off) in response to the bacterial burden. We flagged a bunch of immune-related genes which were differentially expressed in infected and uninfected birds and were just delighted with ourselves.

The paper was published five years ago. Instead  of processing a skip-load of calf thymus to identify and quantify a single molecule that was in vanishingly small quantities [we're almost all water, after all]; we dissected out a lump of wet tissue from a single chicken and measured the concentration of some thousands of genes. That scaling up has been exposed as crude, medieval, against today's standards. A study in last week's Nature has analysed the genetic repertoire of 50,000 single cells from the intestinal epithelium! In my immunological day, we could process white blood cells with an instrument called a flow cytometer. This would flick cells left and right and count them according to what proteins poked out of the cell membrane. The current study un-glues the cells lining the gut and measures each one's genetic activity. That's a level of precision that is truly astonishing, which is why it is published in Nature,  but will look completely normal and even expected in 5 years time.  All those genes in all those cells makes this an example of Big Science: far too much information to read, let alone tally up or assess. without the help of some hefty compute power. The picture [R] is not a Rand-McNally map of the counties of Antarctica; it is rather the first step in the process of making sense of the gut-cell genes. The X and Y axes are the two most informative trends in a multivariate statistical analysis, the diagram shows that known cell-types - goblet cells (island to the L), brush cells (smaller island R) - are not only similar in shape and function (we knew that, it's why they have been given identifying names) but also have very similar genetic profiles.

But it turns out that we made some mistakes when lumping things in the same bin just because they looked similar. We've made this sort of classificatory error before with giraffes - rhinos - elephants and other species. Well it turns out that brush cells aka tuft cells are really two quite different cell types that have rather different gene expression profile [see R pale and dark] and probably therefore have rather different function although it's hard to tell them apart under the microscope. Tuft-1 seems to own neuronal development while Tuft-2 are more down with inflammation. That explains why these two quite disparate functions were associated with 'tuft cells'. We'd be wrong if we classified people-with-helmets as essentially the same: policemen and fire-fighters and motor-cycle couriers and construction workers do different things: the only thing they have in common is an occupational need to protect the head. Tuft-2 cells are all lit up with CD45 which had previously only been associated withe the T- and B-cells of the immune system.

This would be enough for normal people to rest on their laurels but there's much more. And you may bet that, after the flag-ship Nature paper there will be a fleet of filler-of-detail and explore-new-avenue papers over the next tuthree years. Here, as a taster, they show a quite different response to infection by the bacterium Salmonella and the nematode worm Heligmosomoides polygyrus. Whatever about the gene up-regulation profile, it seems that the complement of epithelial cells changes significantly under the two different assaults. Salmonella induces the development of more Paneth cells which then blurf out anti-microbial peptides; H. polygyrus otoh triggers the growth of goblet- and tuft-cells which mount an appropriate response to nematode infection. I think it shows that we would have been 'at nothing' with our blunt instrument approach to different infections. Treating great lumps of tissue as 'essentially the same' and measuring anything about their function rides rough-shod over the subtle cell-by-cell difference in response that is essential to the maintenance of homeostasis against pathogens. It's like you put the entire contents of the Etihad stadium, spectators, astro-turf, goal-posts and all into a blender. . . and are then surprised that you cannot identify the importance of the football.

DNR

Hello? What's not to understand!?  How much clearer do I have to be while unconscious? To a hammer everything looks like a nail; to an ER physician every body cries for intubation and CPR. Almost every time, the person in scrubs wins the dispute. The story, presented as case notes in the NEJM, and picked up by George Dvorsky at Gizmodo and then liked by Neatorama, is that an old chap presented at a Florida ER department a) unresponsive and b) with a signed DNR tatooed on his chest. The ER team intervened aNNyway, because that's what doctors do: decisions, especially those which result in direct action, become [as in are fitting to, but also as in create] a successful medico. But they also had the humility to send a note to the hospital ethics committee. The ethics committee decided that the tattoo was a clearly expressed wish and that it was silly - and wrong - to search for the next-of-kin especially as the old chap presented with no ID.  When my friend and mentor Lynn Margulis had a stroke, her daughter had to fight to get the medical team to heed her mother's clearly expressed terminal care wishes. NEJM

I've been to end of life issues before, and before that, and written a living will; but I haven't gotten the dnr tat yet. They're very expensive, tattoos, maybe I should get a permanent marker and draw a DIY DNR. That way, if I get "tattoo regret" - yes it's a thing - then I just have to scrub regularly or wait to exfoliate. The current advice is don't call the ambulance when you or your aged loved one has their last crisis. If you do, the professionals will do what they do intervening to save lives, even if if means breaking ribs during CPR or stuffing tubes in every orifice and making new holes to take yet more tubes. It is a living for the paramedics, nurses and doctors but it can be ghastly for the Principal on the gurney and eye-wateringly wincing for the attended loved ones. The trouble is knowing which is the last crisis, because most of us would like to see Mum back in her own home patched up for another while.  The last death - at home, in bed, a blessing - I was involved in, the family had to push back quite hard to prevent the corpse being taken to hospital to be pronounced dead. Most deaths occur in hospital and outsiders (GPs) are getting out of the habit of writing death certificates. You can't get the locus or the Care Doc to sign off, it must be the deceased own doctor. Shucks, nobody says that death is easy but it's going to happen for sure; so let's not make it more difficult than it needs to be.