Basic Examples

Get a Dataset containing rows for each sequence:

In[1]:=

Out[1]=

Return the latest date a sequence was included:

In[2]:=

Out[2]=

Count proteins by the reported description:

In[3]:=

Out[3]=

Most of these protein sequences are collected from humans, but not all:

In[4]:=

Out[4]=

Scope & Additional Elements

Get a date plot of collection dates:

In[5]:=

Out[5]=

See a data histogram of release dates:

In[6]:=

Out[6]=

See a timeline plot of inclusion dates:

In[7]:=

Out[7]=

Show the locations where the sequences were gathered:

In[8]:=

Out[8]=

Visualizations

Most of the provided protein sequences come from the United States and Australia:

In[9]:=

Out[9]=

When we look at the geographic locations providing protein sequences with the most common title, "nucleocapsid phosphoprotein", these proportions are largely maintained:

In[10]:=

Out[10]=

Most of the provided sequences come from Victoria, Australia and the states of the Unites States:

In[11]:=

Out[11]=

When we look at the detailed geographic locations providing protein sequences with the most common title, "nucleocapsid phosphoprotein", these proportions are again largely maintained:

In[12]:=

Out[12]=

Analysis

In this data, different submitting parties may title the same protein differently. By gathering all of the titles that share a sequence, we can see that "spike glycoprotein" and "surface glycoprotein" are sometimes used to label the same protein, while the 'structural' designation is applied to matrix, envelope, and nucleocapsid proteins:

In[13]:=

Out[13]=

We can find unique candidate proteins by finding the most common name of the most common sequence per name and eliminating partial and chain sequences:

In[14]:=

Out[14]=

Given these common sequences, we can plot where these sequences are found along the reference SARS-CoV-2 genome. To properly find an alignment, we align with sequences with the translation of each potential frame shift and choose the best alignment:

In[15]:=

Out[23]=