Using the Taxonomy database: Difference between revisions

Latest revision as of 21:45, 27 August 2026

Exercise written by: Rasmus Wernersson and Henrik Nielsen.

UNDER CONSTRUCTION!!!!

Background

When comparing DNA and protein sequences from different species it is important to keep in mind that all living organisms at some point in time has shared a common ancestor. Some organisms are closely related and have recently derived from a common ancestor (e.g. Human and Chimpanzee, which diverged 5-10 million years ago) and some are more distantly related (e.g. Human and mouse, diverged 100-150 million years ago).

The more closely related two organisms are, the more similar their sequences will be (say, when comparing the Alpha Globin gene from each of the organisms), and the more likely it will be that similar looking genes from each organism still have the same function (MUCH more about this when we get to pairwise alignment and BLAST searches).

Apes taxonomy (Detailed taxonomy of the Great Apes: Human, Chimp, Gorilla, Orangutan - from Wikipedia)

Phylogeny vs. taxonomy

As we discussed in the lecture all life is organised in a hierarchical taxonomical system, which approximates the "true" underlying phylogeny to a large degree. It's therefore often important to know where a specific organism is placed in the taxonomical system - this type of information will also always be included along with DNA/Protein sequences from the big databases such as GenBank and UniProt.

Today we will explore various ways to look up and compare taxonomy.

A word about Wikipedia

The free online encyclopedia Wikipedia (and other similar resources) is a GREAT way to start out when you need to look up information about a new topic - in this case taxonomy. Almost all species entries in Wikipedia has a "Scientific classification" box which includes taxonomical information (for example see the entry on Orangutan or Fucus vesiculosus (Bladder wrack / Blæretang)).

HOWEVER: Keep in mind that Wikipedia is NOT a reliable source of information, even if most entries are of a very good quality. The facts in the Wikipedia entries have not been verified by taxonomy experts and can potentially be wrong (everybody can go in and edit the text). We need to look up the taxonomy in an official database (in this case we'll be using NCBI Taxonomy) before you can state it as a fact.

You CANNOT quote Wikipedia as the only source of your information - you'll need to find the original primary source of the information or look it up in an official database.

A word about AI

As with Wikipedia using AI to generate a taxonomical analysis (e.g. comparing how a bunch of species are related) can be a great way to get an overview, and a (typically) well written explanation. You will need to ask the AI to include references to actual scientific sources to document the validity of the results, and you will be responsible for double checking the AI output and making sure not only the data is correct, but also that the conclusions are sound.

The need for a Ground Truth

Luckily, modern taxonomy is very well established and there an internationally recognized organization that makes revisions in a highly regulated manner. This taxonomy standard is captured in the NCBI Taxonomy database we'll be working with in the sections below. For all molecular data NCBI Taxonomy is THE standard reference.

The NCBI Taxonomy Database

Main link: https://www.ncbi.nlm.nih.gov/datasets/taxonomy/

(2026 note: NCBI is in process of migrating to this new site - if you find the old Ncbi Tax homepage via Google, make sure to press the link they provide to jump to the new site).

As mentioned above NCBI Taxonomy will serve as our Ground Truth for everything related to taxonomy. The NCBI Tax database provides the numerical enumeration of species (and other taxonomical levels) that is used and referenced in most Sequence databases, such as GenBank (DNA) and UniProt (Protein). For example human (Homo sapiens) has the ID "9606" and Yeast (Saccharomyces cerevisiae) as the ID "4932".

NCBI Tax is perhaps not a database you would browse for fun (depending on your level of geekiness). It's good for looking up definitions, and for comparing the taxonomical position of multiple organisms (since the information is so densely presented).

Example: Homo sapiens

Open the NCBI Taxonomy webpage in a new browser window/tab (see link above)
Search for "Homo sapiens".
1. Dont Panic: An enormous amount of information is shown - for example about genome sequences. In this case we only need to look at the information presented in the "Taxonomy" tab.
2. Notice the Taxonomy ID - 9606 as mentioned above.
3. "Lineage" (box on the right hand side): Here a condensed overview of the human lineage is shown by default. Notice the taxonomical ranks we talked about in the lecture ("Phylum", "Class" etc). You can navigate to the definition of these groups by clicking on them.
4. "Full lineage": Click this to view a FULL list of all the groups leading "down" to human. Notice, that you can "mouse over" the groups to see a pop-up with the taxonomical rank. NOTICE: The rank "CLADE" is used whenever a group does not have a common English name ("Clade" is the generic name for a uniquely defined taxonomical group).

Play around with the Homo sapiens page for a bit to familiarize you with the interface, and answer the following questions along the way:

QUESTION 1a: What is the TaxID of "Metazoa"?
QUESTION 1b: What is the family that contains humans?
QUESTION 1c: Are humans vertebrates? (Latin: Vertebrata)?

Comparing taxonomy using NCBI Tax

Fruit fly (*Drosophila melanogaster*) - source: Wikipedia

Besides being useful for being the official database behind the TaxID's used in GenBank (and other databases), NCBI Tax actually makes it easy to compare taxonomy.

Let's take the situation where you have read an interesting paper comparing a DNA sequence between the following three organisms: Homo sapiens (Human), Mus musculus (Mouse), and Drosophila melanogaster (Fruit fly), but you have no idea about the relationship between the three organisms.

We can look this up in NCBI Tax:

Open two browser windows/tabs and search for Homo sapiens and Mus musculus.
By comparing the "lineage" text it will be easy to find out at which taxonomical level human and mouse differ.
QUESTION 2A: Look at the (simple/abbreviated) lineage information and find lowest ranking common group for human and mouse - what is the name and what is the rank?
QUESTION 2B: Look at the "full lineage and find the lowest ranking group human and mouse have in common (it's OK if the rank is "clade"). What is the name and TaxId of the group?

Now repeat the analysis by comparing Human and Fruit Fly (Drosophila melanogaster)

QUESTION 3A: What is the lowest rank group shared in the (simplified/abbreviated) lineage? (TaxID, Name)
QUESTION 3B: What is the lowest rank group shared in the full lineage? (TaxID, Name)

Fishing in NCBI Tax using the Common Tree function

Zebrafish (*Danio rerio*) - source: Wikipedia

In this last part of the exercise, we will investigate relationships between different species of fish. We have compiled this list of various fish:

 Latin name             Common name                     TaxID
 Danio rerio            Zebrafish                        7955 
 Gadus morhua           Atlantic cod                     8049 
 Mustelus griseus       Spotless smooth-hound (shark)   89020  
 Petromyzon marinus     Sea lamprey                      7757 
 Latimeria chalumnae    Coelacanth (famous "Blue fish")  7897 
 Lepidosiren paradoxa   South American lungfish          7883

Comparing many species will be tedious by just pointing and clicking, so here we'll utilize the NCBI Common Tree tool.

Now, go to the front page of NCBI Taxonomy and click the Common Tree link a bit down the page (under "Taxonomy resources"). Here, you can add species to a tree one by one by entering either the Latin name or the TaxID in the "search and select" field. (It will try to autocomplete the name you're entering, so pay attention to what is selected when you press enter). You can also add a whole list at once if you have a text file containing either TaxIDs or Latin names, one per line (doable with a bit of editing of the list above).

**Direct link:** Common Tree tool (2026)

IMPORTANT: The 2026 version of the tool is quite good at automatically detecting the shared groups to display, but you can ask the tool to display a full lineage, if needed: You can test this by clicking "View in Taxonomy Browser" after the Common Tree has been populated (but return to the standard view for answering the questions below).

QUESTION 4a: In this selection of species, what is the sister group (nearest neighbour) to the Zebrafish? What is the sister group to the lungfish?

Now try to add yourself (i.e. Human) to the tree, using either the Latin name or the TaxID. Any surprises?

QUESTION 4b: What is now the sister group to the lungfish?

QUESTION 4c: Which of the following is most closely related to the "Blue fish": the cod, or you?

QUESTION 4d: Which of the following is most closely related to the cod: the shark, or you?

QUESTION 4e: Which of the following is most closely related to the shark: the lamprey, or you?

QUESTION 4f: Does the category "fish" make any scientific sense?

Leave the browser window with the Common tree open for the next question.

Comparing trees

A bioinformatician has compared the sequences of a gene from the seven species we used in the previous question, and arrived at the following tree:

You will later learn how to make trees like these in the Phylogenetic trees exercise. For now, you only need to know that such a tree is not necessarily 100% correct, since it is based on a limited amount of data.

QUESTION 5a: Are there any differences in the branching pattern between the gene tree and the Common tree from the previous question?

QUESTION 5b: Can the gene tree be made to comply with the Common tree by swapping two species? If so, which two?

AI investigation

As the final step, use your favorite AI (ChatGPT etc) to try to get as detailed and comprehensive an answer to QUESTION 4f as possible: Does the category "fish" make any scientific sense?

We do realize that the AI field is rapidly evolving, and answers will vary. We will recommend you to play around with the prompt to make the system generate an answer with the best possible scientific rigor. You are welcome to either use the list of species (+human) as we did above, or ask the AI to find relevant data itself.

QUESTION 6: Provide your final prompt + answer from the AI (+ info about which version you used). Reflect upon the correctness of the answer (how much do you trust it, anything fishy?)

@@ Line 40: / Line 40: @@
 ===The need for a Ground Truth===
-Luckily, modern taxonomy is very well established and there an internationally recognized organization that makes revisions in a highly regulated manner. This '''taxonomy standard''' is captured in '''NCBI Taxonomy''' database we'll be working with in the sections below. For all molecular data NCBI Taxonomy is THE standard reference.
+Luckily, modern taxonomy is very well established and there an internationally recognized organization that makes revisions in a highly regulated manner. This '''taxonomy standard''' is captured in the '''NCBI Taxonomy''' database we'll be working with in the sections below. For all molecular data NCBI Taxonomy is THE standard reference.
 ==The NCBI Taxonomy Database==
 [[File:NcbiTax2026.jpg|800px|center|border]]
-In the final part of the exercise we will be using the NCBI Taxonomy database. NCBI Tax is a more "dry" and technical database, which contains accurate (standardized!) hierarchical taxonomy of around 180.000 organisms. The NCBI Tax database also provides the numerical enumeration of species (and other taxonomical levels) that is used and referenced in most Sequence databases, such as GenBank (DNA) and UniProt (Protein). For example human (''Homo sapiens'') has the ID "'''9606'''" and Yeast (''Saccharomyces cerevisiae'') as the ID "'''4932'''".
-NCBI Tax is not a database you would browse for fun (as you might with ToL). It's good for looking up definitions, and for comparing the taxonomical position of multiple organisms (since the information is so densely presented).
+'''Main link:''' https://www.ncbi.nlm.nih.gov/datasets/taxonomy/
+(''2026 note: NCBI is in process of migrating to this new site - if you find the old Ncbi Tax homepage via Google, make sure to press the link they provide to jump to the new site'').
+----
+As mentioned above NCBI Taxonomy will serve as our '''Ground Truth''' for everything related to taxonomy. The NCBI Tax database provides the numerical enumeration of species (and other taxonomical levels) that is used and referenced in most Sequence databases, such as GenBank (DNA) and UniProt (Protein). For example human (''Homo sapiens'') has the ID "'''9606'''" and Yeast (''Saccharomyces cerevisiae'') as the ID "'''4932'''".
+NCBI Tax is perhaps not a database you would browse for fun (depending on your level of geekiness). It's good for looking up definitions, and for comparing the taxonomical position of multiple organisms (since the information is so densely presented).
 ===Example: Homo sapiens===
-# Open the NCBI Taxonomy webpage in a new browser window/tab: http://www.ncbi.nlm.nih.gov/Taxonomy/
+# Open the NCBI Taxonomy webpage in a new browser window/tab (see link above)
 # Search for "''Homo sapiens''".
-## Dont Panic: An enormous amount of information is shown - for example about genome sequences. In this case we only need to look at the Taxonomical information presented at the top of the page. If you panic, take a look at this picture: [[Media:linageguide.png|Link]]
+## Dont Panic: An enormous amount of information is shown - for example about genome sequences. In this case we only need to look at the information presented in the "'''Taxonomy'''" tab.
 ## Notice the Taxonomy ID - '''9606''' as mentioned above.
-## "Lineage": Here the entire hierarchical taxonomy is presented as a densely written list. Each taxonomical group on this list can be clicked and further investigated: first an overview page also containing 3 levels of sub-groups will be shown - click the name again to get to the page dedicated to that entry.
+## "'''Lineage'''" (box on the right hand side): Here a condensed overview of the human lineage is shown by default. Notice the taxonomical ranks we talked about in the lecture ("Phylum", "Class" etc). You can navigate to the definition of these groups by clicking on them.
-#* Notice that a large number of taxonomical groups are listed (including many "in-between" levels such as "''Craniata''" (sub-phylum) and "''Gnathostomata''" (super-class). Most of these groups are simply left out in ToL for brevity.
+## "'''Full lineage'''": Click this to view a FULL list of all the groups leading "down" to human. Notice, that you can "mouse over" the groups to see a pop-up with the taxonomical rank. NOTICE: The rank "CLADE" is used whenever a group does not have a common English name ("Clade" is the generic name for a uniquely defined taxonomical group).
-#* You can switch between showing ALL subgroups and a more condensed list by clicking "'''Lineage'''" - it will switch between "'''full'''" and "'''abbreviated'''".
-#* '''IMPORTANT''': You can investigate the taxonomical rank of any group with out leaving the page by "mousing over" the text.
+Play around with the Homo sapiens page for a bit to familiarize you with the interface, and answer the following questions along the way:
-* '''QUESTION 4''': What is the TaxID of "''Metazoa''"?
+* '''QUESTION 1a''': What is the TaxID of "''Metazoa''"?
-**You can see the group of each level by holding the mouse over the name.
+* '''QUESTION 1b''': What is the family that contains humans?
-* '''QUESTION 5''': What is the family that contains humans?
+* '''QUESTION 1c''': Are humans vertebrates? (Latin: Vertebrata)?
 ===Comparing taxonomy using NCBI Tax===
@@ Line 68: / Line 75: @@
 We can look this up in NCBI Tax:
-# Open two browser windows/tabs (http://www.ncbi.nlm.nih.gov/Taxonomy/) and search for '''Homo sapiens''' and '''Mus musculus'''.
+# Open two browser windows/tabs and search for '''Homo sapiens''' and '''Mus musculus'''.
 # By comparing the "lineage" text it will be easy to find out at which taxonomical level human and mouse differ.
-# '''QUESTION 6''': Turn on "abbreviated" lineage information and find lowest ranking common group for human and mouse - what is the name and what is the rank?
+# '''QUESTION 2A''': Look at the (simple/abbreviated) lineage information and find lowest ranking common group for human and mouse - what is the name and what is the rank?
-# In order to get more information than just a latin name and a taxonomy rank, you can try to look up the group in a different database, such as ToL (NCBI will not reveal more than "placentals" if you investigate it further).
+# '''QUESTION 2B''': Look at the "'''full lineage''' and find the lowest ranking group human and mouse have in common (it's OK if the rank is "clade"). What is the name and TaxId of the group?
-# NOTICE: Since a "user friendly" database such as Tree of Life doesn't contain the same amount of taxonomical groups, it may be necessary pick a group with higher rank if the first one is not found.
-# Open a new browser window/tab and find the information for the Fruit fly.
+Now repeat the analysis by comparing Human and Fruit Fly (''Drosophila melanogaster'')
-* Remember to turn on "abbreviated" lineage information for easy comparison.
+* '''QUESTION 3A''': What is the lowest rank group shared in the (simplified/abbreviated) lineage? (TaxID, Name)
+* '''QUESTION 3B''': What is the lowest rank group shared in the full lineage? (TaxID, Name)
-* '''QUESTION 7''': Which ranked group do connect Human and Fruit fly (ignore "no rank" groups)? Which rank? (You can look up this group in ToL for finding out more information).
 <br style="clear: both" />
@@ Line 94: / Line 101: @@
   Lepidosiren paradoxa   South American lungfish          7883
 </pre>
+<br style="clear: both" />
+Comparing many species will be tedious by just pointing and clicking, so here we'll utilize the NCBI '''Common Tree''' tool.
+Now, go to [https://www.ncbi.nlm.nih.gov/datasets/taxonomy/ the front page of NCBI Taxonomy] and click the '''Common Tree''' link a bit down the page (under "Taxonomy resources"). Here, you can add species to a tree one by one by entering either the Latin name or the TaxID in the "'''search and select'''" field. (''It will try to autocomplete the name you're entering, so pay attention to what is selected when you press enter''). You can also add a whole list at once if you have a text file containing ''either'' TaxIDs ''or'' Latin names, one per line (doable with a bit of editing of the list above).
-Now, go to [https://www.ncbi.nlm.nih.gov/taxonomy the front page of NCBI Taxonomy] and click <u>Common Tree</u> under "Taxonomy Tools". Here, you can add species to a tree one by one by entering either the Latin name or the TaxID in the field near the top and clicking <u>Add</u>. You can also add a whole list at once if you have a text file containing ''either'' TaxIDs ''or'' Latin names, one per line (you might want to try this using block selection).
+[[Image:NcbiCommonTree2026.jpg |400px|frame|center|'''Direct link:''' [https://www.ncbi.nlm.nih.gov/datasets/taxonomy/common-tree/ Common Tree tool (2026)] ]]
-'''IMPORTANT:''' Tick the box labeled "include unranked (phylogenetic) taxa" to get a maximally resolved tree. Note that you can click the small boxes with a "+" to see the full lineages, if you want.
+'''IMPORTANT:''' The 2026 version of the tool is quite good at automatically detecting the shared groups to display, but you can ask the tool to display a full lineage, if needed: You can test this by clicking "'''View in Taxonomy Browser'''" after the Common Tree has been populated (but '''return to the standard view''' for answering the questions below).
-'''QUESTION 8a:''' In this selection of species, what is the sister group (nearest neighbour) to the Zebrafish? What is the sister group to the lungfish?
+'''QUESTION 4a:''' In this selection of species, what is the sister group (nearest neighbour) to the Zebrafish? What is the sister group to the lungfish?
 Now try to add yourself (i.e. Human) to the tree, using either the Latin name or the TaxID. Any surprises?
-'''QUESTION 8b:''' What is now the sister group to the lungfish?
+'''QUESTION 4b:''' What is now the sister group to the lungfish?
-'''QUESTION 8c:''' Which of the following is most closely related to the "Blue fish": the cod, or you?
+'''QUESTION 4c:''' Which of the following is most closely related to the "Blue fish": the cod, or you?
-'''QUESTION 8d:''' Which of the following is most closely related to the cod: the shark, or you?
+'''QUESTION 4d:''' Which of the following is most closely related to the cod: the shark, or you?
-'''QUESTION 8e:''' Which of the following is most closely related to the shark: the lamprey, or you?
+'''QUESTION 4e:''' Which of the following is most closely related to the shark: the lamprey, or you?
-'''QUESTION 8f:''' Does the category "fish" make any scientific sense?
+'''QUESTION 4f:''' Does the category "fish" make any scientific sense?
 Leave the browser window with the Common tree open for the next question.
@@ Line 122: / Line 134: @@
 You will later learn how to make trees like these in the [[Exercise: Phylogeny|Phylogenetic trees exercise]]. For now, you only need to know that such a tree is not necessarily 100% correct, since it is based on a limited amount of data.
-'''QUESTION 9a:''' Are there any differences in the branching pattern between the gene tree and the Common tree from the previous question?
+'''QUESTION 5a:''' Are there any differences in the branching pattern between the gene tree and the Common tree from the previous question?
+'''QUESTION 5b:''' Can the gene tree be made to comply with the Common tree by swapping two species? If so, which two?
+-----
+===AI investigation===
+As the final step, use your favorite AI (ChatGPT etc) to try to get as '''detailed and comprehensive''' an answer to QUESTION 4f as possible: ''Does the category "fish" make any scientific sense?''
+We do realize that the AI field is rapidly evolving, and answers will vary. We will recommend you to play around with the prompt to make the system generate an answer with the best possible scientific rigor. You are welcome to either use the list of species (+human) as we did above, or ask the AI to find relevant data itself.
-'''QUESTION 9b:''' Can the gene tree be made to comply with the Common tree by swapping two species? If so, which two?
+*'''QUESTION 6:''' Provide your final prompt + answer from the AI (+ info about which version you used). Reflect upon the correctness of the answer (how much do you trust it, ''anything fishy?'')