Friday, December 19, 2014

Snarkiness -Or-

100 words on Why I love this Manuscript

Unsupported suppositions
Incomprehensible hypotheses
To our mind
Over-interpreted p-values:
If it’s statistically significant, it must be important
105 words when 10 will do.
The related statistics do not mean much.
Throw it all into a sack without controlling whether it belongs there.
Long meandering sentences with conjunctions that make no sense and verbs so far from their subjects you have no idea what is connected to what : always interrupted with a colon and frequently also with parentheses for self-referring references.*
We have the biggest database we know about.
Erroneous conclusions.

* I can’t even write that kind of sentence when I’m trying!

Monday, December 15, 2014

Friday, December 12, 2014

Nijmegen 2

The 16th A-T Workshop will be held in Beijing next fall. I went to my first ATW, the 5th in the series, in 1993 as a graduate student. 22 years later they're still having them, they're still announcing progress, and they're still looking for a cure. Family-run funding organisations have come and gone. Parents and patients are continually hopeful and disappointed.. One said I gave you millions - where's my cure? Forever elusive.
A couple of treatment options are tentatively on the table, though neither is a cure. Both carry risks. Neither can be tested in a randomized study. People don't agree about going forward with either one. Nor do they agree it is better to be careful.
As often happens, people start leaving before the last session. Trains and planes to catch now or wait a whole day. I'm waiting a whole day, mostly on purpose. This way I have the afternoon to see the town. Not to mention I hate meetings where the last speaker talks to an empty room.
I also hate travelling to a place for a meeting, going straight to the hotel, getting to know the venue really well, and then flying straight out again. Last night's little excursion notwithstanding.

Nijmegen? Oh, I've been there. Really nifty hotel out in the Berg en Dal region.

After a quick lunch I get my camera and check out the 90 minute trail around the park and agricultural area near the hotel. There are several viewpoints marked, but you can't see much from them. The hills aren't very steep and the trees still have most of their leaves. Makes for a nice view of the trees, but they get in the way of any view over the plain.
Now for a bus into town. The oldest part is off-limits to cars, making several blocks of pedestrian-district filled with shops and restaurants and today a long series of outdoor stalls along the main road means the central road is elbow to elbow with people. Throngs. Most of the architecture has been updated, and as I mentioned the large majority of shops are international chains. I'd like to pick up some unique souvenir, something that to me will say Here, without being a chunk of airport schlock made in China. I come across a place selling teas and coffess, with a gorgeous but limited and extremely expensive selection of hand-made pottery. Beautiful stuff but it will be difficult to get home intact. Perhaps I'll stop back for a trinket if I don't see anything else.

Ah, here's Hema. Load up on munchies for home and dutch cookies to share at work. By the time I come out of Hema it's getting dark and the open-air stalls are beginning to pack up. Hey, guys, it's not even 5. Ever thought of outdoor lights?
It starts to rain so I duck into one of the brasseries on the main street for a glass of wine and a cup of spicy tomato soup. Lots of people stop in for an afternoon beer or even dinner.
When I go out again the rain has stopped, the stands are in the last stages of packing, and the streets are deserted. The larger stores are still open, but nobody much is in there but the employees tidying up.
Saturday evening (not even - it's just late afternoon still, according to most of us) and Nijmegen is dead. Maybe they've all gone home for a rest and a shower before coming out for a wild night. Maybe.
I just don't really see that, though.
Several blocks later I catch a bus back to the hotel.

After a scalding hot shower I spend a quiet evening with a book.
At the desk they told we there were trains to Schiphol at 8:42 and at 9:12 that should get me there in time for my flight. To catch the train there are busses at 8:19 and 8:49.
Earlier is better, you know, just in case.
As usual I'm quick to get ready in the morning. I hang around the lobby interminably, then go out into the drizzle to the bus stop. The two blocks don't take nearly as long as they should, so I have plenty of time to hang out at the bus shelter in the dark, pacing against the cold. The rain is weak but makes a lot of nouse in the trees. Leaves come down as drops hit them just right. A day or two of this and they might be naked. The sky lightens gradually, and by the time the bus arrives it might be daytime. Hard to tell. Certainly it is by the time we pull up at the train station.
I have just four minutes to buy a ticket, find my train and get on it. For natives this might be child's play, but my dutch is rudementary. Good thing I have enough margin to take the later train.
Aha. Train to Amsterdam at 8:42, just like the desk clerk said; there's only one leaving at that time, and I find it & hop on just before it starts rolling.
Looking at the display of stops, though, this train does not seem to stop at Schiphol. Um. I suppose trains back to Schiphol from Amsterdam Central must be frequent and quick. Though I know that is a 20 minute trip - good thing I got the early train.
I ask the conductor, and he's all irritated I didn't stamp my ticket in the machine at the station. Tells me I need to change trains at Utrecht. And stamp my ticket there. Be sure to stamp.
OK, ok.
Off the train at Utrecht, there's the Schiphol one marked for quay 7. This is 5. To the right, quays 1-4. To the left, 8-12. Seven, please. Neither right not left leaves the middle, and the train at the opposite quay, not quay 6 but indeed 7 (what do they have against 6? there are other even-numbered quays), is just leaving... Ah, there is another train farther down quay 7, and it is the right one. Leaving immediately.
Ooh. Good thing I decided not to try and get a coffee.
I haven't completely figured out Dutch trains yet. Very helpfully, a screen on the train from Nijmegen told us what the stops where and when we would make them. Close to a stop they display what connections are leaving soon. I was relieved to see my destination, and doubly so on seeing a time 40 minutes later. If I'm going to spend 40 minutes in Utrecht waiting for my connection, thank goodness I took the early train. On the later one I'd miss my plane, or be so stressed about it I might as well. That's why I thought I had plenty of time to get a coffee. But when I saw the train, and the imminent departure time I was lucky to hop aboard in time. Was 40 minutes for the next one? Or did 10:30 refer to our arrival time at the airport (and we do arrive at 10:30) Bit of a difference.

Everything was fine. Nobody looked at my unstamped ticket. The airline was not on strike. There are multiple Starbucks at the airport, on both sides of the security check.

Wednesday, December 10, 2014


Un jour, c’est comme si un couteau a été retiré de la plaie. On se sent tellement bien. On peut commencer à cicatriser, finalement.
Puis le lendemain,
Toujours là.

Ne me dit pas que tu me tiendras courant de quelque chose que tu n’as pas la moindre intention de faire.

Tuesday, December 9, 2014

False economies

So how’s the stagiaire, you might ask, that girl with the unfortunate timing and worse training ?

We were going to look at the clinical data for her samples before deciding which of them to choose to analyse, but ran into the problem that she doesn’t have with her, and cannot get access to, the clinical data (the f*? can’t someone email her a list? – is there no list?) So we have to be satisfied with the labels already assigned.

The problem of how to pick which ones to analyse has been very simply resolved. It turns out that most of the 300+ samples she brought with her are completely degraded and cannot ever be analysed. Only about 60 actually contain DNA in fragments long enough to do anything with. So we’ll go with the 60, whatever they are.

300 blood draws, and 240 of them are good for the trash? What happens when you don’t make the necessary investments up front, but try to get by as cheaply as possible. Instead of using some modern but relatively expensive kit for extracting the DNA, they used the tried and true phenol-chloroform method.

When I was in grad school, a wise post-doc always said “The fast way is the slow way, and the slow way is the fast way.” Very good advice, that. Same can be true of expensive and cheap, when cheap puts success at risk.

Now, you can get acceptable DNA with this method, but you have to really know what you’re doing. This is the method I used way back when I was a tech, and it was not always a success. The first thing to do was to get your reagents right, and that meant spending all the necessary time under the fume hood buffering the phenol to a neutral pH. No neutral pH, no good DNA. Just forget it. Whatever comes out might look good at first glance, then fall to pieces over the next week or month, leaving nothing.
Once you’ve got your phenol neutralized, and you extract the washed white cells from your blood, recover the aqueous phase, re-extract that twice with chloroform to get the phenol gone – and a third time if needed. Yes, just do it again. Phenol is the kiss of death for any further manipulation. Chloroform just evaporates no problem. Now you can precipitate your clean(ish) DNA with ethanol and just a bit of salt if there isn’t enough already. But don’t spin it down. Nice DNA will form a beautiful, fluffy white cloud, like magic, as soon as the ethanol hits it. Seal the end of a glass pipet (ooo yes, get to play with fire in the lab!), and fish the blob of DNA out (do let the pipet cool before dipping it in – otherwise the ethanol at 60% will ignite). Squeeze the DNA blob like a sponge against the edge of the tube and once the ethanol has evaporated, swish it around in a tube of aqueous buffer and let it resuspend for a couple of days before attempting to measure its concentration.

So our student is working on the samples that can be worked on, and we’ll see on Monday what sort of results can be had. Cross our fingers that once she passes the preliminary amplification step, all the rest is ok. After all, the rest all works off of the product of the first step, not the primary material. Because if there are no results, um, well there really is no plan B.