A First Introduction to NCBI

David R. Nelson rev. Feb. 21, 2001

The first place to go for access to sequence data is NCBI the National Center for Biotechnology Information. There is no more comprehensive site. If you will be using bioinformatics at all in the future, then you should create a bookmark file for your bioinformatics links so they will be easily accessible. I have set my browser home button to the NCBI site and one of my first bookmarks is to the blast server at NCBI. The link to NCBI is NCBI. For convenience you should make a separate browser window for these notes so you can move back and forth between screens. In Netscape select FILE, NEW NAVIGATOR, then paste in the URL for this page so you will have a duplicate.

There are many layers and functions at NCBI and we will not go over them all now. To start, I will give you a brief summary of the sequence content in some of the most useful sections of the database. First, go to the Genbank growth statistics

You can see there are about 11 billion nucleotides in 10 million sequences as of Jan. 16, 2001. The growth rate is very steep, with almost 8 billion bases being added in 2000. The following table is taken from the release notes to release 121 of Genbank December 15, 2000.

2.2.7 Selected Per-Organism Statistics The following table provides the number of entries and bases of DNA/RNA for the twenty most sequenced organisms in Release 121.0 (chloroplast and mitochon- drial sequences not included): Entries Bases Species 3918724 6702881570 Homo sapiens 2456194 1291602139 Mus musculus 166554 487561384 Drosophila melanogaster 181388 242674129 Arabidopsis thaliana 114553 203544197 Caenorhabditis elegans 188993 165539271 Tetraodon nigroviridis 151411 125948974 Oryza sativa 218598 106344366 Rattus norvegicus 159473 71215626 Bos taurus 141802 62817102 Glycine max 104535 50991920 Medicago truncatula 91334 49855996 Trypanosoma brucei 97112 49415566 Lycopersicon esculentum 54328 47639714 Giardia intestinalis 77532 47590936 Strongylocentrotus purpuratus 49938 44522016 Entamoeba histolytica 57779 44489692 Hordeum vulgare 83726 40906902 Danio rerio 77506 36885212 Zea mays 18361 32779082 Saccharomyces cerevisiae
Follow the link to molecular databases from the left frame of the NCBI homepage. This page has links on the left frame to the databases EST, GSS and STS. Click on the EST link. At the bottom center click on the Number of ESTs - dbEST summary by organism. I have copied the entries with more than 50,000 ESTs. ESTs are sequence fragments derived from cDNAs, so they represent genes that are expressed as mRNA.
Summary by Organism - February 9, 2001 Number of public entries: 7,352,227 Homo sapiens (human) 3,169,953 Mus musculus + domesticus (mouse) 1,936,347 Rattus sp. (rat) 263,362 Bos taurus (cattle) 160,371 Glycine max (soybean) 147,859 Drosophila melanogaster (fruit fly) 116,471 Arabidopsis thaliana (thale cress) 112,500 Caenorhabditis elegans (nematode) 109,215 Lycopersicon esculentum (tomato) 107,298 Medicago truncatula (barrel medic) 101,752 Danio rerio (zebrafish) 79,237 Zea mays (maize) 76,069 Oryza sativa (rice) 70,969 Hordeum vulgare (barley) 68,665 Xenopus laevis (African clawed frog) 66,076 Chlamydomonas reinhardtii 64,973 Sus scrofa (pig) 57,060 Triticum aestivum (wheat) 54,701 Sorghum bicolor (sorghum) 51,888
Notice that mouse and human have over 5 million ESTs between them. You may be surprised to see that soybean has more ESTs that Drosophila, C. elegans or Arabidopsis, the three completed eukaryote model organisms.
Now go back and follow the link to the GSS database. Scroll down the page to the release notes for the current version and click.
dbGSS release 020901 Summary by Organism - February 9, 2001 Number of public entries: 2,211,010 Homo sapiens (human) 866,507 Mus musculus 612,512 Tetraodon nigroviridis 188,963 Oryza sativa (rice) 93,107 Trypanosoma brucei 90,540 Strongylocentrotus purpuratus 76,019 Arabidopsis thaliana 61,265 Entamoeba histolytica 49,129 Drosophila melanogaster 44,787 Fugu rubripes 42,896 Magnaporthe grisea (rice blast fungus) 12,674 Trypanosoma cruzi 12,245 Leishmania major 11,929
Here are the data for the top organism represented in the GSS database. This is a database of genomic fragments about the same size as ESTs, but these of course contain introns and non-coding regions as well as exons. Please note the organisms represented are different than the ESTs. This can be of help sometimes. The third most abundant organism is Tetraodon nigroviridis, the freshwater pufferfish. These fish genes can be surprisingly similar to human and they are helpful when trying to assemble genes from genomic DNA when you cannot find the intron-exon boundaries. Do not forget them, they can help you.

Below is the same data for the STS section. You can get to it by going back and clicking on STS from the databases page. There are not as many STS fragments. These are genomic sequences made for mapping and most of them have been placed on genetic maps. They are also a last resort if your gene is missing a small fragment and you cannot find this piece anywhere else. It might be here. Keep in mind the human genome sequence as it is now has 170,000 contigs and so there are that many gaps in the sequence. You may need a small piece to fill in a hole in your gene.
dbSTS release 020901 Summary by Organism - February 9, 2001 Number of public entries: 93,728 Homo sapiens (human) 69,963 Rattus norvegicus (Norway rat) 8,361 Danio rerio (zebrafish) 6,036 Drosophila melanogaster (fruit fly) 3,203 Bos taurus (cattle) 1,109 Plasmodium falciparum (malaria parasite) 869 Mus musculus (house mouse) 701 Kluyveromyces lactis 658 Gallus gallus (chicken) 630 Oryza sativa (rice) 339 Sus scrofa (pig) 212 Zea mays (maize) 212 Cryptosporidium parvum 161 Oreochromis niloticus (Nile tilapia) 161

That gives you a flavor of what is in Genbank. Now how do you search this data? The method is the BLAST search. NCBI has recently redone its BLAST search interface. Please go to the new BLAST page from the NCBI homepage. You will notice that there are several types of blast search methods depending on what you have and what you want to find. Notice the first three sections on this page. Nucleotide blast searches, Protein blast searches and Translated blast searches. If you have nucleotide sequence that is from an untranslated region, or you do not know what is in it the nucleotide blast Blastn is what you should try to identify your sequence. It may match a known gene. If you have protein sequence and you want to search the known protein translations in Genbank BlastP is one of your choices. The translated blast searches are more powerful and they take more computer time. Tblastn is one of the most useful search tools. This is the one I use almost all of the time with fragments of genes that I have been able to translate into probable open reading frames based on other known protein sequences. With Tblastn you can search your protein sequence against all the nucleotides in Genbank sections that are translated in all six reading frames (three per DNA strand). You can hardly miss anything that matches if it is out there. Nucleotide searches are not as sensitive because there are only four bases but there are 20 amino acids, so there are better search statistics to find a match. The Blastx program lets you find ORFs (open reading frames) in your nucleotide sequence by hunting for matches to translated proteins. This is very helpful in analyzing fresh genomic sequence that is unannotated. The Tblastx program is the most computer intensive since it does the translation of the query sequence in six frames as well as the translation of the nulceotide database in 6 frames, so it is like doing 6 tblastn searches. Not too many servers will even offer this search, since it eats computer time so badly.

The Blast server page has a window to paste your sequence. This may be in FASTA format or just raw sequence. FASTA format always starts with an >IDENTIFIER LINE This may be anything you want to identify your sequence. This will appear in the search output. If you will be doing many searches and saving them or printing them, it is a good idea to include this identifier line so you will not mix up your output. The > symbol is required, otherwise the program will try to interpret your identifier as amino acid or nucleotide sequence. The sequence can have numbers in it such as are found in the Genbank sequence format, or other formats like GCG or PIR. These will be ignored. Non-standard letters like O, and J will also be ignored. For this example use this sequence.

>Dictyostelium CYP51 P450 C-terminal ETQKDINDIVQKENQGEINFDGLKRMNRLETVIREVLRLHPPLIFLMRK VMTPMEYKGKTIPAGHILAVSPQVGMRLPTVYKNPDSFEPKRFDVED KTPFSFIAFGGGKHGCPGENFGILQIKTIWTVLSTKYNLEVGPVPPTD FTSLVAGPKGPCMVKYSKKQK*

Once your sequence is in the window, you need to select a program, like tblastn from the pull down menu. Choose Tblastn. You also need to select a database from another pull down menu. Select nr (the default). We will talk about the other options later. Below the sequence window is the blast button. Do not click this yet. Futher down the screen is the filter option. I usually recommend turning the filter off, unless there is a weird repetive sequence like a run of QQQQQ in your sequence. If you leave the filter on, you may find some unpleasant runs of XXXXXXX in the blast output where you do not want it. Now one of the best feratures of this blast page is the limitation of sequences to be searched. Here you can use the pull down menu to select a common species like mouse (Mus musculus) or a whole taxonomic range of organisms like green plants (Viridiplantae) or Fungi. I really wish they had some more of these ranges incorporated here, but you can also choose to type in a more specific set yourself like Diplomonadida or Rhodophyta. I use this feature all the time to restrict the search output to just what I want. You may select Dictyostelium discoideum from the pull down menu if you want to find the exact sequence you are searching with, or you may expand the search to closely related organisms by typing in Mycetozoa. I think it may be more interesting to use Fungi on this first blast, since there are many CYP51s known from fungi in the nr database. Go ahead and click the blast button.

You will get a screen that says Format on a button. Click this button to begin formatting your results. This will give you another window that says how long the process will take before the next refresh. This could be short or long depending on usage at NCBI. Monday mornings are bad. Sometimes they say it will be 60 minutes or more. Evenings it may turn around in less than a minute. The output will have a graphical box at the top that has colored bars representing the hits. Red is for the highest scoring hit and the colors get cooler as the scores drop. Red hits are generally pretty good matches. Since we searched a slime mold vs. Fungi here, the best hits are magenta not red. Below the graphics box is a text list of your hits with the best scores at the top. The e-19 numbers refer to the chance that this was an accidental match. The larger the negative exponent the better the match and the lower the probability that it was a chance occurrence. Below the text list are the actual alignments. This is where you see what you have found. The output here shows a gap in the Dicty sequence compared to most of the fungal sequences. Since fungi have relatively few introns, this is probably not an intron in the fungi, but it may be an insertion that is found in fungi and not in the slime mold sequence. Notice that all the top hits are identified as CYP51 sequences or 14 alpha demethylase enzymes. This is because the cytochrome P450 family is extensively annotated and over 1000 of the genes are named. Even the new sequences that are not officially named are usually identified by this type of blast search and the authors often tag them as a P450 (in this case CYP51).

Go to Module 2