Projects

ANTIBODY BINDING PLOTS AND 4PL OR 5PL CURVE FITTING (AbPlot)
MOLECULAR BIOLOGY SUITE (MolBioWiz)
TEXT EXTRACTION TOOL (Tet)
AMINO-ACID COLORING IN MICROSOFT WORD (AaColoringVBAmacro)
POSITIONAL RESIDUE FREQUENCY STATISTICS (AbAAposStats.py AbAAposStatsPlot.py)

ANTIBODY BINDING PLOTS AND 4PL OR 5PL CURVE FITTING (AbPlot)↑

Purpose: Develop a web-based graphing program to plot and analyze antibody binding data.

Development: AbPlot uses the the plotly.js library. I coded the user interface and a least squares regression function to fit the data to the 5PL equation in HTML/JS/CSS.

Features: The program takes up to 12 sets of antibody binding data, which are ideally copied from an Excel or Google Sheet and pasted into a text area. The data are automatically fitted into a 5PL equation and plotted on a semi-log scatter chart.

y = R + \frac{L - R}{{(1 + {(\frac{x}{k})}^{S})}^{s E}}

The 5PL parameters can be individually selected for re-fitting and re-plotting. For example, one can freeze the E parameter with the value of 1, which effectively converts the equation into 4PL, or set the plateaus to the known minimum and maximum, while re-fitting the data to the remaining parameters. The axes can be set to linear, or log and ticks can be either automatically or manually set. Colors, marker shapes and line types can also be individually selected. This first version does not calculate and plot error bars.

Note: I have written a similar program in Python, which while not being as convenient to use as this this web-based version, could be more versatile for Python users. It is available from GitHub: AbBindingCurves.py Here are some figures created by the python program:

MOLECULAR BIOLOGY SUITE (MolBioWiz)↑

Purpose: Develop a web-based expandable suite of programs for molecular biology applications, and especially for antibody engineering.

Development: I wrote MolBioWiz in HTML/JS/CSS.

Features:

TEXT functions

grep extracts lines with (regex) keywords.
findAll outputs all the found (regex) keyword positions.
replace creates a copy of the text with (regex) keyword.
substring extracts a column of text between two indices.

SEQUENCE functions

all SEQ functions can process thousands of fastA-formatted sequences.

XYZ removes non-sequence characters and capitalizes all characters.
X-Y removes only blank spaces.
RC reverse-complements sequences.
wordSize and lineSize formats the sequence output.
show res #s shows residue numbers in the sequence output.
fastA⇄name::seq toggles between fastA and single line formats.
transl translates nucleotide sequences.

from a 5' motif
from a 5' motif plus n nucleotides.
in translation frames 1, 2, 3, in any combination.
displays DNA sequence over translation.

name edits or changes sequence names.

padding normalizes the length of all sequence names.
replace global (regex) find/replace in sequence names.
splice splices out parts of names between indices.

find extracts sequences that have a (regex) pattern.

PROTEIN functions

fetchFromUniprot fetches protein sequences from a list of accession numbers.
Properties calculates protein properties from their sequences.

ANTIBODY functions

Abalign aligns antibody variable region sequence(s) to the imgt numbering system.
Developability find sequence motifs that might cause Ab developability problems.

PEPTIDE function

pepScan finds peptides to use as immunogens.

This figure shows a typical Abalign output followed by the automatic detection of sequences prone to deamination in the CDR regions.

TEXT EXTRACTION TOOL (Tet)↑

Purpose: The Unix program grep finds and outputs all text lines that contain a keyword, which can be a regular expression. The purpose here was to write a program like grep that extracts additional relevant text lines.

Development: I wrote Tet in HTML/JS/CSS.

Features: There are four areas in the user interface:

FILES: select hundreds of files from your computer to be parsed. Files with extensions other than .txt .text and .csv will be ignored.
KEYWORD: enter the keyword or regex in this area and chose options.

case-sensitive.
extract n lines before or after after the line with the keyword.
extract the sentence that contains the keyword even if it spans more than one line.

TEXT: Parse a single file, which may be uploaded here. Alternatively, one can copy and paste text here, including from non- .txt .text .csv files.
OUTPUT: the output of the search is sent into this area.

Tet

AMINO-ACID COLORING IN MICROSOFT WORD ↑

Purpose: Write a macro in VBA to color amino-acid residues in multiple-sequence alignments in Microsoft Word documents, resulting in colored text that is fully editable.

Development: The VBA code AaColoringVBAmacro is available from GitHub. In the VIEW - MACROS tab of an MS-Word document create a new macro and paste the code into it.

Features: The advantage of using this code over those of most software is that you get a Microsoft-Word file, which is fully editable, as opposed to an image. The alignment on the following figure was generated with molBioWiz and colored in Word using this VBA macro.

POSITIONAL RESIDUE FREQUENCY STATISTICS ↑

Purpose: Write Python programs to gather statistics of individual amino-acid residue frequencies at specific sequence/structure locations. This kind of data is a requirement for any protein engineering project.

Development: The Python scripts are available from GitHub. AbAAposStats.py AbAAposStatsPlot.py.

Features: The scripts, respectively calculate and plot the frequency of each amino-acid residue at each position of a multiple sequence alignment.
As an example, I generated multiple sequence alignments of, ~2,500 human and ~500 human rabbit Ig-VH-FR3 sequences taken from the Kabat database. For this, I used my molBioWiz program. Next, I used the Python scripts to generate the amino-acid frequency data and to plot it.