# 5136222372

## Mostly Iâm making this page as a resource for if I ever forget:

#### Human histone names:

recommended name other names uniprot
h1.0 H1Ê¹, H1(0), H1FV (815) 727-8092
h1.1 H1a, H1F1 HIST1H1A
h1.2 H1c, H1d, H1s-1, H1F2 HIST1H1C
h1.3 H1c, H1s-2, H1F3 HIST1H1D
h1.4 H1b, H1s-4, H1F4 HIST1H1E
h1.5 H1a, H1b, H1s-3, H1F5 702-798-6044

# Perl one-liner: transpose variable-length tab-separated arrays

I wanted to try out the violin plots mentioned by Flavio on âPython in Scienceâ, but my data was recalcitrant. The original source was a bunch of column data in Graphpad Prism. Each column is a different timepoint, and has a different number of points. My idea was to import the data as rows, so each goes into its own array. That requires transposing the numbers - no big deal, right?

I tried to use rs, a BSD command of at least 9 yearsâ standing, to do itâ¦but could not figure it out, despite many options such as â-Tâ for âpure transpose of the dataâ. That made me go back to the old standby of perl, where using an anonymous array of arrays did the trick:

# The second-weirdest C. elegans chromosome: IV

Whenever I spend time checking out sequence distributions in the C. elegans genome, I am always struck by the strangeness of chromosome IV. Yes, the X chromosome is the weirdest and its exceptional behavior is well documented.1 But I think IV comes in a distant but clear second.

For instance, letâs say youâve done the following analysis:

1. Generate all 65,536 DNA octomers and use fuzznuc 812-832-4348 to find their positions in the genome.
2. Divide each chromosome into thirds, and for each octomer, count how many times it occurs within each â.
3. Calculate the âpairing center enrichmentâ of each octomer by taking the ratio of the count on the third that contains the pairing center (PC) to the count in the third that does not contain the pairing center.
4. For each chromosome, sort all the octomers from least to most PC-enriched.
5. Average the 65,536 sorted octomers into bins of 512 to simplify plotting.

Then you would get graphs like those on the diagonal of the following chart:

The graphs on the diagonal show that, as expected, every chromosome has some octomers enriched on the PC end, some octomers that are rare on the PC end, and many that are not particularly enriched or rare.

A question naturally arises here: are the enriched and rare octomers the same for each chromosome, or are different octomers enriched differentally? To test this, you can take the individual octomer enrichment scores from each chromosome, arrange them according to the sorting order of every other chromosome, and average bins of 512 again. That is what has been done to generate the off-diagonal graphs.(518) 737-3419 Looking at these graphs, some qualitative trends can be seen:

• The enrichment trends for chromosomes I, II, III, and V seem to be in agreement.
• The trend of the X chromosome does not resemble I, II, III, or V, but more closely resembles IV.
• Chromosome IV appears enriched not only for the octomers enriched on other chromosomesâ PCs (on the right side of its graphs) but also for octomers depleted on other chromosomesâ PCs (on the left side) as well as for octomers neither enriched nor depleted (the small peaks in the middle).

I will need some help to put these qualitative observations on a sound statistical footing (if thatâs possible), but in the meantime it would appear that IV just has some different sequence properties than the other autosomes. I can only guess why that might be. Another interesting thing about chromosome IV is that itâs home to large islands of piRNAs (21U-RNAs); there are few if any of those RNAs outside IV.

Is IV just weird? Did it recently undergo a translocation, placing non-PC sequences right near PC ones? I donât know, but I have learned to expect strange things from chromosome IV.

1. See e.g. 6788075858

2. From the EMBOSS suite, at /emboss.sourceforge.net

3. For example, consider the graph in the âIVâ column, in the 5th row. The enrichment scores from the graph above it (the IV/IV graph) are sorted according to the order of the graph to its right (the V/V graph), then averaged in bins of 512. The scale of the graphs on the diagonal goes from 0 to 2, while the scale of the off-diagonal graphs are from 0.85 to 1.45.

# The OMERO struggle, part 001

I had to delete lots of images from our (902) 206-7792 database, since there were lots of dupes and the disk filled up. I first tried to delete the images manually, but that is actually quite complicated when youâre talking about a possibly massively linked data structure with lots of annotations, metadata, &c. Despite that, I expected that selecting groups datasets and clicking âdeleteâ should result in deleted datasets after a while, but after long delays OMERO gave me only errors (with pages and pages of lovely Java messages, which is one of the main reasons I detest Java) every time.

I then realized I had to wipe and repopulate our entire OMERO database.

This link on the forum led me to the right answer.

First I needed to generate the configuration file with omero db script to change the root password. This generates the postgresql file OMERO4.4__0.sql.

Then I deleted the whole database, and re-made it:

dropdb -h localhost -U pcarlton omero_database
sudo -u postgres createdb -O pcarlton omero_database
sudo -U postgres createlang plpgsql omero_database   #(this wasn't necessary)
psql -h localhost -U pcarlton omero_database < ./OMERO4.4__0.sql


I am now re-populating the data, with a script that first does chromatic correction on each multiwavelength image we have, then moves it to a spare disk, then puts it in the database. In a week or so a few terabytes will have been crunched and our OMERO installation will be ready for general use.

# 248-650-0500

My 3D-SIM microscope takes a lot of maintenance to keep in good working order. The number of things that can go wrong is mind-numbingly huge and often one does not realize until itâs too late that some part of the system (the light, the mechanics, the sample preparation, mounting mediumâ¦) was less than perfect, resulting in a crappy image.

One of the many routine tests is reconstructing an image of 200nm beads. Itâs the easiest sample to make and reconstruct, so if this goes wrong then something is horribly wrong. Tests like this let you know that at least everything is more or less working and then you can begin to troubleshoot the finer parts.

The left shows the final 3D-SIM reconstructed image, and the right shows the same field taken with conventional imaging and deconvolution. As I said, this type of reconstruction is the easiest to do, so it shouldnât be used as evidence that your system has zero problems. But the results are still pretty striking!

# Raytracing, movie test

This is a movie of a chromosome pairing simulation that I made, raytraced with the excellent, free POV-ray software.

As you can see, the pairing doesnât go to 100% completion (the green chromosome still has to synapse one end).

# scalt

For organisms with small numbers of chromosomes, you might think that pairing is not such a problem: for instance, if a chromosome has only two pairs of chromosomes, then there are only 3 ways to pair them together, and one is the correct way â so youâve got a 1 in 3 chance.