Thesis writing reference

[edit] 1 Generals

• Parentheticals discouraged
• Make sure paragraphs stay in the same tense
• Parallel sentence construction
• Information flow goes from what is already known to what is new
• When asserting something is true, go for the first reference. When asserting the state of the field, go for the latest reference.
• Make sure that when using "it" and "this" that it is absolutely obvious what they refer to; it probably isn't, so make sure to specify
• Be very obvious; don't assume information is already known
• In procedure: say why things are being done, then what we did
• Keep citations at the end of sentences
• No dangling participles
• Paragraph: topic sentence, what, conclusion. Start with why, go to what, recap.
• When making a reference, make sure to define whether it is to a figure, section, or table
• Try to use the same term instead of closely related terms (ie, pick one of frequencies or percentages, etc)
• Do not start a sentence with "they"
• Add legends to all graphics; they must stand alone
• Make the first sentences of things in Results AND Discussion fill in gaps for someone who didn't read the methods or the results
• End each paragraph with a take home message for that paragraph
• Decimal numbers like .5 have the leading 0: 0.5

[edit] 2 Specifics

• A gap is the object in the alignment; an indel is the event that causes them
• Evolution should not be personified or imply goals or motive or decision making
• No "third dimensional structure"; three dimensional structure
• Amino acids can NOT be called bases--nucleotides are bases. Use residues when referring to amino acids.
• Phylogenies are not built, they are inferred.
• Phylogeny is a noun. Phylogenetic is the adjective. Thus, it is a phylogenetic tree, not a phylogeny tree.
• Data is plural; data are not data is
• The word "different" is generally redundant

[edit] 3 Consistent terminology

• Referring to tables and figures in parens, not (see Figure X) but (Figure X).
• Frequency over percentage
• Substitution matrix over log odds matrix
• log odds value not log odd value
• Log odds value, not substitution value
• Substitution matrix instead of substitution log odds matrix, which is overkill
• Characterized sequence, NOT characteristic sequence
• Talk about evolution selecting against and not for
• Substitution, not transition
• Don't use band at all for percent identity; use level
• Off-diagonal, not nondiagonal

[edit] 4 Consistent graphics

• Disordered graph bars before ordered graph bars
• Disordered is red, ordered is blue
• 85% lighter than 60% lighter than 40% in side by side bar graph comparisons
• Log odds: negative is red, positive is blue, neutral is lilac
• Log odds: ordered is always subtracted from disordered. Thus, higher in disordered is blue and lower in disordered is red.