Forum:Bioinformatics "Cheat Sheet"

Question

[Chris Miller]

Inspired by Keith Robison's post on 'cheat sheets', what would you put on a cheat sheet for bioinformatics? This might include one-line scripts, conversion factors, handy rules of thumb, etc.

Some of Keith's suggestions, which have a biology slant:

• IUPAC ambiguity codes for nucleotides:
• Amino acid single letter codes.
• SI prefixes in order.
• Powers of 2.
• Tm calculation estimation using G+C and A+T counts.
• 1 human genome ~= 7 pg of DNA
• 1 bp = 660 daltons

[Chris Evelo]

Could you please collect the answers and put them on a cheat sheet blog somewhere?

[Piotr Byzia]

Instead of blog post, maybe github repo with Markdown/LaTeX would be better?

[Michael Schubert]

And/or incorporating answers here would be nice.

[Dataminer]

Hey! Brilliant idea to have a cheat code. but this list will go endless unless you give sub categories, like the cheat code for researchers working in bioalgorithm development, genomics, data analysis etc... this will make it more organised.

Answer 1

[Pierre Lindenbaum]

• 5' : left
• 3' : right

;-)

[Ian]

-1 my apologies to Pierre as my objection is rather pedantic; if you are looking at coordinates relative to the forward strand (e.g. Refgene), then a gene on the reverse strand would be 5' right and 3' left.

[Michael Schubert]

+1 for the smile while reading ;)

[Madelaine Gogol]

I actually have a post it note on my cubicle wall with a little picture of genes on each strand and 5' and 3' with little arrows.

[Pierre Lindenbaum]

@Ian fair enough :-)

Answer 2

[Madelaine Gogol]

Not completely bioinformatics oriented, but some things I've found handy.

#subtract a small file from a bigger file
grep -vf filesmall filebig

#use awk to rearrange columns
awk '{print $2 " " $1}' file.txt

#sort a bed file by chrom, position
sort -k1,1 -k2,2n file.bed > file.sort.bed

#strip header
tail +2 file > file.nh

#find and replace over multiple files
perl -pi -w -e 's/255,165,0/255,69,0/g' *.wig

#print line 83 from a file
sed -n '83p'

#insert a header line
sed -i -e '1itrack name=test type=bedGraph' file.bed

#sum column one from a file
awk '{s+=$1} END {print s}' mydatafile

[Chris Penkett]

One of my awk aliases is the mean and sd of column 1:

awk '{s+=$1;s2+=($1*$1)} END {print s/NR,sqrt((NR*s2-s*s)/(NR*(NR-1)))}'

[Madelaine Gogol]

Oh, that could be useful too, thanks. awk is still a dark interesting rabbit hole to me.

Answer 3

[Aaronquinlan]

I have a vision of this cheat sheet being an extensive, very convenient set of environment variables and man pages. It should be versioned and should be on something like GitHub.

For example:

####################
# HG19
####################
$CHR1_SIZE=249250621
$CHR2_SIZE=243199373
...

####################
# Shortcuts
####################
$SUMCOL='awk '\''{ SUM += $1} END { print SUM}'\'

Other informational stats should be rolled into "man" entries. For example,

man dna
man iupac
man 2_powers
man log_examples

This may be utterly harebrained, but it seems useful to me. A community-based, focused wikipedia and shortcut library on the command line.

[Pierre Lindenbaum]

Hi Aaron, I started this today :-) https://github.com/lindenb/bioman

[Casey Bergman]

+1 for very clever idea. I like this a lot.

[Chris Miller]

Dotfiles can be intensely personal things. That said, I'd love to have a big central repository of useful stuff to pick and choose from.

[Aaronquinlan]

Fair point, yeah a repo that is organized by type would be more useful.

[Ketil]

Eh, shortcuts isn't a cheat sheet, it's a .bashrc file. So:

sumcol(){
   awk '{SUM += $1} END { print $SUM }'
}

But +1 for the manual pages suggestion, one of the man pages I constantly return to is man ascii. Can we create a b section?

Answer 4

[Alastair Kerr]

missing from the list

• 1 nucleosome = 147bp
• Crude AA to kilo dalton conversion = AA No X 0.11 =Kd
• Perl one liners for text conversion
• s/015012/012/ # Windows -> Unix
• s/012/015012/ # Unix -> Windows

[Ying W]

or

perl -pi -e 's/rn/n/g' input.file

Answer 5

[amolkolte1989]

I have a collection of handpicked reference cards. It helps every now and then.

I prefer to call it as a Bioinformatician's Pocket Reference!!

Answer 6

[Fred Fleche]

Correspondance between the genome version nomenclature : hg19 (UCSC) = GRCh37 (NCBI)

[Malachi Griffith]

The UCSC Assembly Releases and Versions FAQ does a great job of summarizing a lot of these. Each genome build in the table lists: species, UCSC version, release date, release name/id, and status.

[Danielk]

...except for chrM/MT where UCSC have a different sequence than the accepted correct one.

[Chen Sun]

Warning: there is a small difference between hg19 and GRCh37 that make a significant influence in the downstream analysis:

in GRCh37, the chromosome name is 1,2,3,4,5,6,7,8,9,..., X, Y

in hg19, the chromosome name is chr1, chr2, chr3, chr4, ..., chrX, chrY

So the mapping results to hg19 cannot be used to GRCh37 directly.

Hope others can avoid the trap I fall in.

[Pierre Lindenbaum]

and some degenerate bases have been replaced by 'N' for chr3 and chrY. see: http://plindenbaum.blogspot.fr/2013/07/g1kv37-vs-hg19.html

Answer 7

[Mary]

This reminds me a little bit of BioNumbers: http://bionumbers.hms.harvard.edu

Answer 8

[Thaman]

• AUG = Initiation

• UAA, UGA, UAG= Termination

• AT= 2 Hydrogen Bond, GC =3 Hygrogen Bond and adjacent bases are separated by 3.4Å

• Purine= Adenine & Guanine AND Pyrimidines= Cytosine, Uracil & Thymine

• DNA replication is semi-conservative

Coming more.... :D

Answer 9

[Michael Schubert]

• Amino acid weights, IEPs
• some FASTA statistics one-liners
• quick overview of possible cli BLAST inputs/outputs (reading -help takes so long as they are all over the place)
• BLAST tabular output column names
• Karlin-Altschul formula
• definition of PAM and BLOSUM
• order of AAs in a substitution matrix/PSSM

Answer 10

[Aleksandr Levchuk]

The cheat sheet for programming in R would be what you are looking for.

Here are good manuals that my advisor, Thomas Girke, wrote:

• R & Bioconductor Manual
• HT Sequence Analysis with R and Bioconductor

The HT Sequence Analysis manual was as recommended in Recommend Your Favorite Introductory "R In Bioinformatics" Books And Resources

Answer 11

[Chris Miller]

I'll start off with a few of my own:

• an alpha-helix has 3.6 residues per turn

• A haploid human genome has a little over 3 billion bases and contains around 20,000 genes

• A handy alias for summing up a column of numbers from the command line:

  sumcol='awk '\''{ SUM += $1} END { print SUM}'\'

Answer 12

[Jeremy Leipzig]

I would like a cheat sheet of arguments for common bioinformatics executables (e.g. blast, clustal, bowtie, bwa, fastx-toolkit), the popular bioperl scripts (like bp_seqfeature_load.pl), as well as the most common bioinformatics things in bash (e.g. mass renaming: foreach f in *fasta; do mv $f `echo $f | sed -e 's/.fasta/.fa'` done)

[Alastair Kerr]

Check out the 'rename' program that come with perl is so much better. In this case

rename 's/fasta$/fa/' *fasta

(I assume it comes with perl as it was written by Larry Wall- it is standard on all the latest ubuntu systems)

[Frédéric Mahé]

With bash, you can use pattern substitution: for f in *.fasta ; do mv $f ${f/fasta/fa} ; done. It is more than twice faster than calling for sed (on a set of 1000 files).

[Neilfws]

I find running the executable without arguments usually reminds me what they are ;-)

[Eric Normandeau]

Mass renaming is fun until you have to do it on someone else's directory and the file names are full of spaces and accents...

Answer 13

[Pierre Lindenbaum]

My cheat sheet would contain the length of the human chromosomes.

[Chris Miller]

For which assembly? :-P

Answer 14

[Kevin]

building up my list here.. a blog post would be a good record for myself when i change computers or move office where i usually lose my printed copies.

http://kevin-gattaca.blogspot.com/2011/03/cheat-sheets-galore-bioinformatics.html

Answer 15

[David W]

Very cool question, here's mine, which probably isn't all that relevant to most biostar members but is popular in our lab:

A table of nucleotide substitution models, and how to set them in the most commonly used programs

Still working on it (you can implement the exotic models in most of the software, but not easily)

Answer 16

[Pals]

My cheat sheet would be

Amino acid structures with their properties

And I would also consider Biostar because it is in fact more than google.

Answer 17

[ALchEmiXt]

A useful addition would be a landscape or flowchart how to get from one file (format) into the next..... bioinformatics is about parsing it right....... :)

[Kevin]

interesting idea .. but off the top of my head I can only think of fastq (1)-> bam (2)-> vcf (3) -> annotated SNPs list of which the path taken depends on the sofware used to (1) map/align (2) call SNPs etc ... Are there file formats that you are thinking about?

[ALchEmiXt]

Maybe some simples are the interconversion of fastq and fasta+qual; fastq (qual solexa) to fastq (sanger and so forth); conversion of annotation files like EMBL, GBK into each other and or gff; conversion of all sorts of IDs (but there are some good tools for that)....and may be some more....

Answer 18

[Neilfws]

I like this question, so at the risk of sounding trite: my cheat sheet = a Google search. I store very little information these days; it's as quick and easy to search for it as and when required.

[Jeremy Leipzig]

that has an uneven success rate. which google query will lead you directly to the answer of, for example, what is the percentage of the human genome contained in transcription units?

[Jeremy Leipzig]

"At present, about one-third of the human genome appears to be transcribed" http://bit.ly/ga2YFU just the amount of surfing I had to do and still not find that number is evidence enough that a genomics cheat sheet would be handy thing

[Chris Evelo]

Did you try this one? "human genome percentage transcribed" It gives this as the first hit: http://bionumbers.hms.harvard.edu/bionumber.aspx?s=y&id=103746&ver=2 That is a nice bonus, but the second hit: http://www.genome.gov/25521554 tells you what you asked for (1.5-2%)

[Jeremy Leipzig]

that has an uneven success rate. which google query will lead you directly to the answer of, for example, what percentage of the human genome contained in transcription units

[Chris Evelo]

Did you try this one? "human genome percentage transcribed" It gives this as the first hit: http://bionumbers.hms.harvard.edu/bionumber.aspx?s=y&id=103746&ver=2

Answer 19

[Paige]

Great ideas! I'd add a Blosum62 substitution matrix to the list.

Answer 20

[Samuel Lampa]

List of most used file formats (.pdb, .bam, .fastq, etc etc), what information they contain, and what they can be used for? (and possibly the most well-known software(s) that reads them)

Answer 21

[Michi]

great idea!

a bit of biology:

the citrus cycle! http://student.ccbcmd.edu/~gkaiser/biotutorials/cellresp/images/u4fg35.jpg

or here you can find it also along other must-knows

for R & Regex I already have separate cheatsheets on my desk. One thing I am missing tough, is a cheatsheet for Regex, referring to in which environment one has to escape which characters and back-references (\ or $)

[Ying W]

i really like this website for regex http://www.sarand.com/td/ref_perl_pattern.html

Answer 22

[Goldbear]

Biology by the numbers

http://www.rpgroup.caltech.edu/publications/SnapShot2010.pdf

Answer 23

[hadasa]

• No. seqs(fasta):

  grep \> file_name | wc -l
  

[Pierre Lindenbaum]

there's even a shorter solution : grep -c ">" filename :-)

[Aleksandr Levchuk]

@Pierre, I always liked the piped version more. It's only 3 or 5 symbols longer but it easily to swap wc with less or another grep à la LEGO.

[brentp]

grep > file_name will just truncate your file. needs quotes around the ">" as per @Pierre.

[hadasa]

the editor seems to escape my > so had to write \>

[Darked89]

grep -c "^>" your_fasta

The > sign has to be the first on the line

Answer 24

[Prakki Rama]

##Tabulated BLAST header
qseqid sseqid pident alignlength mismatch gapopen qstart qend sstart send evalue bitscore **- **

## go to the end of file in Vi editor
G (shift + g )** **

##substitute in Vi editor

:%s/Soxgene/Foxgene/g

##remove exact duplicate sequences from fasta file

sed -e '/^>/s/$/@/' -e 's/^>/#/' file.fasta | tr -d '\n'|tr "#" "\n"| tr "@" "\t" |sort -u -t ' ' -f -k 2,2 |sed '/^$/d'|sed -e 's/^/>/' -e 's/\t/\n/'

##remove blank lines
sed '/^$/d' file.fasta >Noblanks_file.fasta.out